) (05/15/91)
chip@tct.com (Chip Salzenberg) writes: > According to mathew@mantis.co.uk (CNEWS MUST DIE!): > >"Model net.citizens" do not hack the news software to silently discard > >articles which it is perfectly capable of delivering. > > Quite. Which is why G&H have done no such thing. > > True, C News drops articles on the floor. But it logs having done so, > and it notifies the Usenet administrator in the output of newsdaily. So what the hell use is that? None of those Usenet administrators told *ME* about it, and I'm the one who has to fix my software. > True, it is sometimes possible to deliver articles that do not conform > to the relevant RFCs. But if that were done, of what value would the > RFCs be? If an article can be corrected to conform, then it should be, and should then be delivered. If an article cannot be corrected, then either 1) It should be dropped and an attempt made to notify the article originator, or 2) It should be passed on anyway in the hope that the next site will have better software and be able to correct it; and optionally an attempt should be made to notify the article originator. Option 2 might mean that people at offending sites get lots of automatic mail. In which case, good! It'll encourage them to fix their software, and the news will still be being delivered. With the situation as it is at the moment, there are STILL sites who don't know that all their articles are being dropped, so non-conforming articles will continue indefinitely. > Geoff and Henry have done the Right Thing. Rubbish. Throwing away data containing correctable errors, without issuing any warning message, is NEVER the right thing. > Don't shoot the messenger. There was no messenger. C News didn't tell me, remember? > Fix your software. I have, and now I want to get C News fixed. mathew
dylan@ibmpcug.co.uk (Matthew Farwell) (05/15/91)
In article <5ukZ24w164w@mantis.co.uk> CNEWS MUST DIE! <mathew@mantis.co.uk> writes: >chip@tct.com (Chip Salzenberg) writes: >> According to mathew@mantis.co.uk (CNEWS MUST DIE!): >> >"Model net.citizens" do not hack the news software to silently discard >> >articles which it is perfectly capable of delivering. >> Quite. Which is why G&H have done no such thing. >> True, C News drops articles on the floor. But it logs having done so, >> and it notifies the Usenet administrator in the output of newsdaily. >So what the hell use is that? None of those Usenet administrators told *ME* >about it, and I'm the one who has to fix my software. [ As a bit of background, we are mantis's main feed and we are the site that dropped six of their articles ] I did tell you. I have a mail message sitting in my outgoing folder to you and 7 other news administrators who we exchange news with to the effect that cnews now drops articles which are badly formatted. Its dated April 30th 1991. Do you want me to resend it to you? If you are talking about noone telling you about cnews dropping some of your articles, then I also told you about that, within a few hours of me seeing the log message. I sent you back the log entries and told you to resend them. Unfortunately I haven't actually got a copy of that mail message but I distinctly remember sending it. I even have your reply to it somewhere. >> Geoff and Henry have done the Right Thing. >Rubbish. Throwing away data containing correctable errors, without issuing >any warning message, is NEVER the right thing. But it *did* issue a warning message. It was logged. Look, put it this way. You say mail should be sent to the originator if my site can't correct it. Who will pay for this? From July it'll cost me (if I remember correctly) about 7.5p/k for international outgoing mail. A quick grep through my log files for dropped articles reveals that most of them (say 90%) are international addresses. I get an average of 20 dropped articles a day. Taking the average length of a mail message back to the sender to be 1k, this would cost me about 18*7.5 = 1.35 pounds a day ~= 500 pounds/year. Are you going to pay that for me? If not, then who is? I'm not going to. This figure doesn't even include phone bills. >> Don't shoot the messenger. >There was no messenger. C News didn't tell me, remember? Yes, but I did. I warned you in advance. I told you when your articles had been dropped. Dylan. -- Matthew J Farwell: dylan@ibmpcug.co.uk || ...!uunet!ukc!ibmpcug!dylan You can't kill me, I'm a genius!!!!
henry@zoo.toronto.edu (Henry Spencer) (05/16/91)
In article <5ukZ24w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >> Fix your software. > >I have, and now I want to get C News fixed. It is not broken and consequently will not be fixed. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
ghelmer@dsuvax.uucp (Guy Helmer) (05/18/91)
In <5ukZ24w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >chip@tct.com (Chip Salzenberg) writes: >> True, C News drops articles on the floor. But it logs having done so, >> and it notifies the Usenet administrator in the output of newsdaily. >So what the hell use is that? None of those Usenet administrators told *ME* >about it, and I'm the one who has to fix my software. >> True, it is sometimes possible to deliver articles that do not conform >> to the relevant RFCs. But if that were done, of what value would the >> RFCs be? >If an article can be corrected to conform, then it should be, and should then >be delivered. Henry has pointed out many times that it is difficult to correct problems in an article. Direct human intervention is essentially the only way to fix incorrectly formed headers. >If an article cannot be corrected, then either >1) It should be dropped and an attempt made to notify the article originator, >or >2) It should be passed on anyway in the hope that the next site will have > better software and be able to correct it; and optionally an attempt > should be made to notify the article originator. No. You better not pass on the article, or you're part of the problem. Attempts to notify the article originator would be a bad thing to do, too. A site producing news may not receive mail or there may not be a path for a message to follow back to the sender. Worse yet, if you attempt to both pass on the article and notify the sender, you and everyone running software like yours will clog the net with tons, er, megabytes, of useless messages. Should the originator be unreachable via mail, the net will bite it. And what about a message that's had it's headers munged by a single machine running junky software? A person would be deluged with messages from stupid news software incorrectly attributing the bad article to him/her and his/her software. >Option 2 might mean that people at offending sites get lots of automatic >mail. In which case, good! It'll encourage them to fix their software, and >the news will still be being delivered. Bogus, dude. >With the situation as it is at the moment, there are STILL sites who don't >know that all their articles are being dropped, so non-conforming articles >will continue indefinitely. >> Geoff and Henry have done the Right Thing. >Rubbish. Throwing away data containing correctable errors, without issuing >any warning message, is NEVER the right thing. There are warning messages. Be a good admin---correctly configure your machine's news and mail software (i.e., have news send mail about problems to usenet and alias usenet to the news admin id, but don't depend on news to tell you when it is having trouble)---and watch your system. >> Don't shoot the messenger. >There was no messenger. C News didn't tell me, remember? >> Fix your software. >I have, and now I want to get C News fixed. Dump your self-righteous attack against C-news. News isn't mission-critical; news isn't life-sustaining; delivery of articles isn't guaranteed. The RFC's exist, and if your software doesn't implement them correctly, that's tough. Check your software to make sure it generates RFC-compliant articles. That's the only way you can be guaranteed that you have a good chance of getting your article distributed throughout the net. You can't blame anyone or any software for dumping your article and not telling you about it, especially if your software is faulty. >mathew -- Guy Helmer, Dakota State University Computing Services helmer@sdnet.bitnet, dsuvax!ghelmer@wunoc.wustl.edu, ghelmer@dsuvax.dsu.edu "Everybody need a soft filter / Everybody need reverse polarity" - Rush
) (05/18/91)
henry@zoo.toronto.edu (Henry Spencer) writes: > In article <5ukZ24w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) w > >> Fix your software. > > > >I have, and now I want to get C News fixed. > > It is not broken and consequently will not be fixed. It loses articles unnecessarily, without reporting its delivery failure to the originator. I call that broken. If sendmail did that, you'd call it broken, wouldn't you? mathew
rickert@mp.cs.niu.edu (Neil Rickert) (05/18/91)
In article <swT423w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >It loses articles unnecessarily, without reporting its delivery failure to >the originator. > >I call that broken. I don't. Good work, Henry and Geoff, in improving quality. >If sendmail did that, you'd call it broken, wouldn't you? I certainly would. There is, however, an enormous difference between email and news. In principle, email is person to person communication. Getting the message through is therefore the principle. Adherence to exact standards is not as important (as long as this does not cause system problems), as getting the message to its destination. Sender and recipient can have their own private arguments about standards. (No debates about this here, please. This is a forum on news software). News, on the other hand, is public, not private. It is in some sense, a form of publication. Any failure to observe standards is an imposition on the general public, so the publisher (in this case the news software, under supervision of the administrator), has a higher obligation to maintain accepted standards. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
chip@tct.com (Chip Salzenberg) (05/19/91)
According to mathew@mantis.co.uk (CNEWS MUST DIE!): >It loses articles unnecessarily, without reporting its delivery failure to >the originator. I'd disagree with the "unnecessarily" part. >I call that broken. I don't. >If sendmail did that, you'd call it broken, wouldn't you? Sure. But Sendmail doesn't do news. -- Brand X Industries Custodial, Refurbishing and Containment Service: When You Never, Ever Want To See It Again [tm] Chip Salzenberg <chip@tct.com>, <uunet!pdn!tct!chip>
chip@tct.com (Chip Salzenberg) (05/19/91)
According to mathew@mantis.co.uk (CNEWS MUST DIE!): >rdc30@nmrdc1.nmrdc.nnmc.navy.mil (LCDR Michael E. Dobson) writes: >> News software which generates articles that do not conform to the RFC is >> broken. Articles generated by such buggy software deserve to be discarded. > >Well, perhaps you'll explain that to the poor user who posts for weeks into >the void, unaware that the software he's using is generating bad headers. Administrators who get "bogus header" messages and don't pass them on deserve flamage. Software that (finally!) implements the relevant RFCs doesn't. -- Brand X Industries Custodial, Refurbishing and Containment Service: When You Never, Ever Want To See It Again [tm] Chip Salzenberg <chip@tct.com>, <uunet!pdn!tct!chip>
nreadwin@micrognosis.co.uk (Neil Readwin) (05/19/91)
In article <5ukZ24w164w@mantis.co.uk>, CNEWS MUST DIE! <mathew@mantis.co.uk> writes: |> If an article cannot be corrected, then either [...] |> 2) It should be passed on anyway in the hope that the next site will have |> better software and be able to correct it; and optionally an attempt |> should be made to notify the article originator. |> |> Option 2 might mean that people at offending sites get lots of automatic |> mail. In which case, good! Erik Fair's recent experiences with sendmail (described in c.p.tcp-ip I believe) seem to indicate that this is *not* good :-\ Neil. Phone: +44 71 528 8282 E-mail: nreadwin@micrognosis.co.uk Everything is a cause for sorrow that my mind or body has made
ben@wri.com (Ben Cox) (05/19/91)
In <1991May17.170033.17759@dsuvax.uucp> ghelmer@dsuvax.uucp (Guy Helmer) writes: >Henry has pointed out many times that it is difficult to correct >problems in an article. Direct human intervention is essentially >the only way to fix incorrectly formed headers. This is the lazy man's answer. Given sufficient effort, one could probably write a shell script that would fix a Date: header. A Path: header would be harder, but most others are cake. -- Ben Cox ben@wri.com My boss didn't tell me to write this.
henry@zoo.toronto.edu (Henry Spencer) (05/19/91)
In article <swT423w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >> It is not broken and consequently will not be fixed. > >It loses articles unnecessarily, without reporting its delivery failure to >the originator. >I call that broken. You may call it what you wish. Our experience is that all alternatives are worse. The losses, while regrettable, are necessary. They are reported to the local sysadmin, which is the best that can be done safely. If you want this to be changed, propose a workable, fully thought out alternative without serious side effects. (Hint: you're going to have to put some serious effort into this, because all the obvious schemes occurred to us long ago. You needn't bother repeating them, and I'm not going to bother telling you yet again what's wrong with them.) -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
kherron@ms.uky.edu (Kenneth Herron) (05/19/91)
chip@tct.com (Chip Salzenberg) writes: >>> News software which generates articles that do not conform to the RFC is >>> broken. Articles generated by such buggy software deserve to be discarded. >Administrators who get "bogus header" messages and don't pass them on >deserve flamage. Software that (finally!) implements the relevant RFCs >doesn't. One problem with this is that the C-news report only lists the site(s) which are feeding you bad articles, not where the articles actually came from or the exact problem with the article. The message IDs are logged along with the complaint , but the messages themselves are thrown away (of course) so the concerned administrator has to do a bit of detective work. I'm working on a script to notify me of all articles bit-bucketed for bad header reasons. I could arrange for it to be posted, e.g. to news.lists. Is there any call for such a thing? -- Kenneth Herron kherron@ms.uky.edu University of Kentucky +1 606 257 2975 Department of Mathematics "So this won't be a total loss, can you make it so guys get to throw their mothers-in-law in?" "Sure, why not?"
cactus@zardoz.club.cc.cmu.edu (Todd Masco) (05/19/91)
In article <ben.674596383@dragonfly.wri.com> ben@wri.com (Ben Cox) writes: >This is the lazy man's answer. Given sufficient effort, one could probably >write a shell script that would fix a Date: header. Oh? What would you do with a header, Date: 910201 Is it January 2nd or February 1st? Okay, you're going to have to 1) Fudge some things arbitrarily, or 2) Drop some badly formed articles and keep others. I'm not going to touch rewriting headers into a form that may be wrong (IE, Jan 2 becomes Feb 1). I consider that alternative unacceptable. As has been said by othes, there's a reason behind RFCs. Better to drop everything that doesn't conform to the standard -- otherwise, some people will have some articles that mysteriously drop while others posted by the same person will die. "My articles were working," the user will whine to the admins. "What changed?" The admins in at least one case would probably scratch their heads for a few hours and then give up. The standard is almost useless if it isn't conformed to by everybody. Dropping all non-conforming articles makes it perfectly clear what the problem is -- the RFCs aren't being followed. -- Todd L. Masco - CMU Physics | "Your god has gone and from now on, you'll cactus@zardoz.club.cc.cmu.edu | have to learn to hate the things you fear.
henry@zoo.toronto.edu (Henry Spencer) (05/19/91)
In article <1991May19.013616.10567@ms.uky.edu> kherron@ms.uky.edu (Kenneth Herron) writes: >... The message IDs are logged >along with the complaint , but the messages themselves are thrown away >(of course) so the concerned administrator has to do a bit of detective work. This is something we'd kind of like to do *something* about -- it's not nice to throw away the evidence -- but defining exactly what should be done is a little tricky and we haven't yet sorted it all out. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
rickert@mp.cs.niu.edu (Neil Rickert) (05/19/91)
In article <1991May19.013616.10567@ms.uky.edu> kherron@ms.uky.edu (Kenneth Herron) writes: >I'm working on a script to notify me of all articles bit-bucketed for >bad header reasons. I could arrange for it to be posted, e.g. to >news.lists. Is there any call for such a thing? Unless you have enormous numbers of such articles, the attached patch might do part of what you need. It just keeps the whole batch in in.coming/bad . After that it should be easy to run a script to pull the headers from the particular articles, then discard the batch. *** newsrun.dist Sat May 18 22:14:50 1991 --- newsrun Sat May 18 22:15:42 1991 *************** *** 147,153 **** # history entries for refused articles; this is right for # NNTP-feed sites and doesn't hurt uucp-feed sites unless # they refuse a good fraction of what they get. ! relaynews -r -n $stale <$text st=$? if test $st -ne 0 then --- 147,153 ---- # history entries for refused articles; this is right for # NNTP-feed sites and doesn't hurt uucp-feed sites unless # they refuse a good fraction of what they get. ! relaynews -s -r -n $stale <$text st=$? if test $st -ne 0 then -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
jhenders@wimsey.bc.ca (John Henders) (05/19/91)
In article <1991May18.192836.137@micrognosis.co.uk> nreadwin@micrognosis.co.uk (Neil Readwin) writes: >In article <5ukZ24w164w@mantis.co.uk>, CNEWS MUST DIE! ><mathew@mantis.co.uk> writes: > >Erik Fair's recent experiences with sendmail (described in c.p.tcp-ip I >believe) seem to indicate that this is *not* good :-\ Neil. > Eric's problem arose from sendmail attempting to deliver the article, as well as notifying him of the problem. If the first C-news site that rejected the article dropped it on the floor, this problem wouldn't exist. I think an attempt to notify the sender is a good idea. It may not always get through,but at least it tried. Matthew, Mail me if you want patch info for the mmail program. John Henders Vancouver,B.C.
nreadwin@micrognosis.co.uk (Neil Readwin) (05/20/91)
In article <1991May19.102741.24411@wimsey.bc.ca>, jhenders@wimsey.bc.ca (John Henders) writes: |> Eric's problem arose from sendmail attempting to deliver the article, |> as well as notifying him of the problem. Indeed, and if you read what mathew wrote that is *exactly* what he was suggesting C news should do. He also said "Option 2 might mean that people at offending sites get lots of automatic mail. In which case, good!". I think he underestimates the impact that 'lots of automatic mail' can have. Futhermore it is not obvious that dropping the article and sending mail actually solves this problem given the connectivity of non-C news sites that *could* continue to propagate the article. Phone: +44 71 528 8282 E-mail: nreadwin@micrognosis.co.uk Everything is a cause for sorrow that my mind or body has made
tp@mccall.com (Terry Poot) (05/20/91)
In article <1991May18.211109.20401@zoo.toronto.edu>, henry@zoo.toronto.edu (Henry Spencer) writes: >You may call it what you wish. Our experience is that all alternatives >are >worse. The losses, while regrettable, are necessary. They are >reported >to the local sysadmin, which is the best that can be done safely. If >you >want this to be changed, propose a workable, fully thought out >alternative >without serious side effects. (Hint: you're going to have to put >some >serious effort into this, because all the obvious schemes occurred to >us >long ago. You needn't bother repeating them, and I'm not going to >bother >telling you yet again what's wrong with them.) Are you saying that this is a problem that doesn't merit attention? I just found out from 2 sysadmins (one in Kentuky, USA, one in New Zealand), that C news is now dropping my articles on the floor. I don't keep track of C news, I was unaware that there was a problem with the newsreader I use, the subject line of this thread is not exactly informative (I don't know why I actually read any of it, I've been ignoring it for some time), and C news reporting of the problem appears to be obscure enough that only 2 sysadmins on the entire net were willing to put forth the effort to determine where the messages came from, and who sent them. This could have been done with prior announcement, an attempt to find non-conforming news software and notify users and authors of the impending change, and some mechanism to ease the transition (like generating a specific report that would list for each article specifically affected by the change, the From: line, Message-Id:, and what was wrong with the article. Sysadmins that care could then do something about it. You could have even done something reasonable, like in this version bitching about the formatting problem, but accepting it, and in the NEXT version, rejecting it. That would have given the rest of the net some chance fix our software BEFORE you started throwing away articles. In article <28357D59.17BB@tct.com>, chip@tct.com (Chip Salzenberg) writes: >Administrators who get "bogus header" messages and don't pass them on >deserve flamage. Software that (finally!) implements the relevant RFCs >doesn't. Everyone other than the 2 people who told me about the problem, consider yourselves flamed if you have my articles in your bogus header file. I am reasonable sure that there are more the 2 entry-points in to otherwise fully connected C news sites on this net. I'd say hundreds of sysadmins out there are ignoring bogus header messages as a matter of routine. Chip, if you aren't one of them, I'd say your neighbors must be, because if you didn't get my articles, your neighbors dumped them. I'd guess very few sysadmins are doing anything at all about these messages. In article <1991May19.022322.29056@zoo.toronto.edu>, henry@zoo.toronto.edu (Henry Spencer) writes: >In article <1991May19.013616.10567@ms.uky.edu> kherron@ms.uky.edu >(Kenneth Herron) writes: >>... The message IDs are logged >>along with the complaint , but the messages themselves are thrown away >>>(of course) so the concerned administrator has to do a bit of >detective work. > >This is something we'd kind of like to do *something* about -- it's >not >nice to throw away the evidence -- but defining exactly what should be >done >is a little tricky and we haven't yet sorted it all out. Maybe you should have thought it out a little further before releasing this version? In article <1991May19.022219.18956@cs.cmu.edu>, cactus@zardoz.club.cc.cmu.edu (Todd Masco) writes: >In article <ben.674596383@dragonfly.wri.com> ben@wri.com (Ben Cox) >writes: >>This is the lazy man's answer. Given sufficient effort, one could >probably >>write a shell script that would fix a Date: header. > >Oh? What would you do with a header, > >Date: 910201 You are aware of some software currently in use that generates this? >... >As has been said by othes, there's a reason behind RFCs. Better to >drop everything that doesn't conform to the standard -- otherwise, >some people will have some articles that mysteriously drop while >others posted by the same person will die. "My articles were >working," the user will whine to the admins. "What changed?" The >admins in at least one case would probably scratch their heads for a >few hours and then give up. > >The standard is almost useless if it isn't conformed to by everybody. >Dropping all non-conforming articles makes it perfectly clear what the >problem is -- the RFCs aren't being followed. Have you ever heard of a defacto standard? The defacto standard for dates in usenet messages has for sometime been that set of dates recognized by B news, which has apparently been more generous than the RFC. The C news change now drops messages that have always been acceptable to every news system on the net. The authors are correct in that we should comply with the RFC, but it was handled very poorly. I, as a reader of news.admin, should have been aware of this change before it was propagated world wide, installed, and started junking my messages, which are generated by dxrn, and are acceptable to ANU News, C news (all previous versions), B news, and to the best of my knowlege, all other news software in use on the net. Obviously, Henry Spencer is under no obligation to do anything at all in particular with software he writes, so don't even start that flame war. He doesn't have to even know standards exist. However, please don't try to construe this action as being in the public good. It wasn't. It could have been, if transition had been considered. -- Terry Poot <tp@mccall.com> The McCall Pattern Company (uucp: ...!rutgers!ksuvax1!deimos!mccall!tp) 615 McCall Road (800)255-2762, in KS (913)776-4041 Manhattan, KS 66502, USA
) (05/20/91)
ghelmer@dsuvax.uucp (Guy Helmer) writes: > In <5ukZ24w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: > >chip@tct.com (Chip Salzenberg) writes: > >> True, C News drops articles on the floor. But it logs having done so, > >> and it notifies the Usenet administrator in the output of newsdaily. > > >So what the hell use is that? None of those Usenet administrators told *ME* > >about it, and I'm the one who has to fix my software. [ Point not answered ] > >If an article can be corrected to conform, then it should be, and should the > >be delivered. > > Henry has pointed out many times that it is difficult to correct > problems in an article. And I have pointed out that certainly in this specific case (my articles being thrown away due to a bad Date: header) it is trivially easy to fix the header. Mail systems do it. > No. You better not pass on the article, or you're part of the problem. If the error is non-critical and you can still pass on the article, why not pass it on? Software which falls over in a heap as a result of bad input is extremely broken, and C News should not attempt to Nanny it. > Attempts to notify the article originator would be a bad thing to do, too. > A site producing news may not receive mail or there may not be a path > for a message to follow back to the sender. In which case the notification will fail to get through. Fair enough. Mail bounce messages sometimes fail to get through, but that doesn't mean you shouldn't try to send them. > Worse yet, if you attempt to both pass on the article and notify the sender, > you and everyone running software like yours will clog the net with tons, er, > megabytes, of useless messages. Fair enough. So either pass on the article or notify the sender. I'm happy with either. All I'm not happy with is not passing on the article AND not notifying the sender. > And what about a message that's had it's headers munged by a single machine > running junky software? A person would be deluged with messages from > stupid news software incorrectly attributing the bad article to him/her > and his/her software. Deluged with one message, assuming that you don't both forward and report. Or maybe ten or so messages if you're in a bit of the net which is exceptionally well-connected via non-C-news sites. > >Rubbish. Throwing away data containing correctable errors, without issuing > >any warning message, is NEVER the right thing. > > There are warning messages. Be a good admin---correctly configure your > machine's news and mail software (i.e., have news send mail about problems > to usenet and alias usenet to the news admin id, but don't depend on news > to tell you when it is having trouble)---and watch your system. I do. I read our system logs, I check for mail bounces and assist, I respond to queries, I check out errors. Unfortunately, there isn't a log file on our system which will tell me whether UUNET's copy of C News is throwing away all my articles. If there were, there would be no problem. ALL I AM ASKING is that if a posting from a site X is dropped, SOMEONE from site X should be told about it. This notification should NOT rely on some random system administrator happening to take pity on site X. > Dump your self-righteous attack against C-news. News isn't mission-critical; > news isn't life-sustaining; delivery of articles isn't guaranteed. True. However, delivery of articles is much less guaranteed now that C News works in "drop all data on error" mode. As to news not being important -- some people spend literally hours writing stuff to post to Usenet. I'd rather that they continued to do so. And they won't if they think that some big C News site will throw it all away because they made a typo in a header field. > The RFC's > exist, and if your software doesn't implement them correctly, that's tough. Interesting attitude. Am I to take it that you would not object if I wrote a program to forge cancel messages for all articles which were not STRICTLY RFC compliant? > Check your software to make sure it generates RFC-compliant articles. That's > the only way you can be guaranteed that you have a good chance of getting > your article distributed throughout the net. So what if I'm an arts graduate? Do you really expect every random user of Usenet to read through the RFC documents and check the news software which exists on his system? > You can't blame anyone or > any software for dumping your article and not telling you about it, > especially if your software is faulty. Of course I can. C News did not need to dump my article. It deliberately chose to do so, however. It also deliberately chose not to warn me, and so the error continued. Therefore it was at least partly C News' fault that the articles were lost. mathew
) (05/20/91)
henry@zoo.toronto.edu (Henry Spencer) writes: > In article <swT423w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) w > >> It is not broken and consequently will not be fixed. > > > >It loses articles unnecessarily, without reporting its delivery failure to > >the originator. > >I call that broken. > > You may call it what you wish. Our experience is that all alternatives are > worse. Why, which have you tried? > The losses, while regrettable, are necessary. Why is it "necessary" to drop articles with the date written as Date: Mon 20 May 91 15:46 GMT instead of Date: Mon, 20 May 91 15:46 GMT then? Enquiring minds want to know. And why is it "necessary" to drop articles with a header line such as Keywords:news,errors or Summary: (blank, with no space after the ':') then? > They are reported > to the local sysadmin, which is the best that can be done safely. If you > want this to be changed, propose a workable, fully thought out alternative > without serious side effects. Outline algorithm for dealing with errors: a. If error can be corrected, correct it. b. If error cannot be corrected and involves the message-ID, drop the article and report failure. c. If error cannot be corrected and does not involved the message-ID, pass on the article. This is an idealized algorithm. In practise, I don't care if you drop all erroneous articles, so long as a failure report gets back to the originating site. Possible algorithms for reporting failure: 1. Mail to originator, or to Errors-To: address if one is specified. If the return address is not mailable, hard luck -- but at least we tried. 2. Mail to administrator of site which produced bad article. If sitename is mangled in return address, hard luck -- but at least we tried. 3. Send a report to a central error-reporting machine (say, UUNET). The central machine merges reports caused by articles with the same message-ID, and sends back a single error report to someone at the guilty site. 4. Propogate failure reports using the news software mechanisms. Alongside "control" and "junk" have an "errors" newsgroup. This at least makes sure that error reports get back to the source site. The message-ID can be generated in a fixed manner from the bad article's message-ID, and so duplicate error reports will be eliminated with virtually no impact on bandwidth. Alternatively, use "control" to report errors. This has the advantage that you don't need to worry about getting existing sites to carry the "errors" newsgroup. > (Hint: you're going to have to put some > serious effort into this, because all the obvious schemes occurred to us > long ago. Well, the above schemes are the ones I can come up with after about five minutes of not-too-painful thinking. I'd call them all "obvious". Number four is my current favourite, and I've not seen any argument which invalidates it. > You needn't bother repeating them, and I'm not going to bother > telling you yet again what's wrong with them.) Sorry? You're saying that: (a) You're not going to tell me what alternatives you've considered, not even by mailing me a file listing them; (b) That you've thought of all the obvious schemes; and (c) That if one of my schemes is obvious, you're not going to bother saying what's wrong with it but just reject it? I hate to inject a note of personal criticism, given that your fan club is already sending me hate mail, but isn't that just a little bit smug and self-satisfied? mathew
) (05/20/91)
chip@tct.com (Chip Salzenberg) writes: > Administrators who get "bogus header" messages and don't pass them on > deserve flamage. Unfortunately, it seems there are quite a few sites out there whose system administrators have better things to do with their time than read the log files and send mail messages to people whose articles have been junked. As a general principle, it is unwise to design software in such a way that it relies on system administrators being conscientious individuals. mathew
emv@msen.com (Ed Vielmetti) (05/21/91)
In article <1991May20.093018@mccall.com> tp@mccall.com (Terry Poot) writes: >The standard is almost useless if it isn't conformed to by everybody. >Dropping all non-conforming articles makes it perfectly clear what the >problem is -- the RFCs aren't being followed. Have you ever heard of a defacto standard? The defacto standard for dates in usenet messages has for sometime been that set of dates recognized by B news, which has apparently been more generous than the RFC. The C news change now drops messages that have always been acceptable to every news system on the net. One of the reasons for no longer accepting and trying to parse mangled dates is that the date parser needed to cope with the multitude of slightly off date formats is complicated and slow. C News has much faster date munching routines, but they don't understand every possible conceivable date that there is. The slowness of the old date parser was why C News didn't check that articles were too old and thus passed along the periodic bursts of old news. The law of unintended consequences says that whenever you fix one problem, you get a new one. In this case, the tradeoff is between passing old news (bad for everyone) vs. dropping malformed articles (bad for a few individuals). Given some time, everyone who currently has software that puts out mangled dates will be properly coercing into fixing them, and the net as a whole will be more happy. --Ed
billd@fps.com (Bill Davidson) (05/21/91)
In article <1991May20.093018@mccall.com> tp@mccall.com (Terry Poot) writes: >C news is now dropping my articles on the floor. I don't keep track of >C news, I was unaware that there was a problem with the newsreader I use, That's your problem. It was announced in advance. (The plans have been on display at the local planning office at Alpha Centauri for ... ;-). Henry and Geoff have been threatening for a long time and they finally made good on their threats. When the patches were put out, Henry posted a big warning message that things were different now. >C news reporting of the problem appears to be obscure enough that only 2 >sysadmins on the entire net were willing to put forth the effort to >determine where the messages came from, and who sent them. What makes you think your articles even got to very many admins? It's very possible that you don't have a non-Cnews path to the entire net. At some point there may have been only to places for your article to continue and both were Cnews sites. I'ts very posssible and even probably though I don't know much about your section of th net. Your path to me indicates a couple of KSU sites. Are they running Cnews? >This could have been done with prior announcement, It was. >an attempt to find >non-conforming news software and notify users and authors of the impending >change, How do you propose to do that? Do you have even the faintest clue as to how many systems are on the net or how many different types and versions of software generate or pass news articles? Don't think about it. It will make your brain hurt. >and some mechanism to ease the transition (like generating a >specific report that would list for each article specifically affected by >the change, the From: line, Message-Id:, and what was wrong with the >article. Sysadmins that care could then do something about it. >You could have even done something reasonable, like in this version bitching >about the formatting problem, but accepting it, and in the NEXT version, >rejecting it. That would have given the rest of the net some chance fix our >software BEFORE you started throwing away articles. People have been bitching about bad news software for years. Even Bnews bitches in the log file about bad headers. I suspect Cnews did too, even when it didn't drop articles. A lot of people don't fix their software until forced to. There are people out there running Bnews 2.9. There may even be older versions out there but that was the first version that respsonded to the version control message. >Obviously, Henry Spencer is under no obligation to do anything at all in >particular with software he writes, so don't even start that flame war. He >doesn't have to even know standards exist. However, please don't try to >construe this action as being in the public good. It wasn't. It could have >been, if transition had been considered. It was for the public good. News software that generates bad headers needs to be fixed or replaced. People have been begging for it. News software bitches about the lack of it. Henry warned everybody about it. I'm glad that I get less bad header messages in my log file and I don't care that I miss a few articles from sites that have bad news posting software. In fact I'm glad. USENET addiction wastes too much of my time anyway (I know, that's MY problem ;-). --Bill Davidson
tp@mccall.com (Terry Poot) (05/21/91)
In article <EMV.91May20135904@crane.aa.ox.com>, emv@msen.com (Ed Vielmetti) writes: >one problem, you get a new one. In this case, the tradeoff is between >passing old news (bad for everyone) vs. dropping malformed articles >(bad for a few individuals). Given some time, everyone who currently >has software that puts out mangled dates will be properly coercing >into fixing them, and the net as a whole will be more happy. I don't have a problem with the change. I have a problem with the shock treatment theory of getting people to fix old software (~drop their news on the floor, that'll get 'em to fix their software!~ [~ instead of " because it isn't a quote]. Why couldn't this have been phased in? Say have it log the error for a few months, so that people could be notified that they had software that needed fixing, THEN start dropping it on the floor. Also, many people have said that it could log these better. That should have been a big consideration, since SOMEONE has to tell sites that C news doesn't like their headers. BTW, the mail has already started rolling in. In response to a few points raised via email: It has been claimed that this was announced thoroughly in advance. If this is the case, I apologize for that complaint. However, I don't read articles that look like they are about C news, because I don't run it. I don't read articles about bogus headers, because my software works (you can't redefine something valid as now being bogus, and expect people to know this before they read the message). Given these considerations, if I did indeed miss something I shouldn't have, I apologize for complaining about it. I only read this thread because it has been going for a while, and I was curious. I figured it was probably about rec.humor. I was informed that most site admins didn't tell me C news was dropping my messages, because they didn't want to overburden me. Dear site admins, you are being WAY too conservative. I've only gotten 2 messages, both in just the last few days. There are probably lots of people out there not being informed at all that they have a problem. BTW, I don't need any more notifications. I've fixed my software. It was a 10 minute fix, and I'd have happily done it before C news was changed, had I known. I've sent it to the author, and one of the 2 of us will post the patches, probably by tomorrow morning. But please warn other people that have the problem. Don't assume someone else will. -- Terry Poot <tp@mccall.com> The McCall Pattern Company (uucp: ...!rutgers!ksuvax1!deimos!mccall!tp) 615 McCall Road (800)255-2762, in KS (913)776-4041 Manhattan, KS 66502, USA
mjr@hussar.dco.dec.com (Marcus J. Ranum) (05/21/91)
While I'm usually in 100% agreement with Henry Spencer about what constitutes the Right Way to do software, I'd like to make one little observation on the subject of dates: We PAY FOR COMPUTERS TO DO THE GUNCH WORK that humans don't want to - such as scanning and parsing dates - that's what they are there for. I personally don't care if my server spends 5% more time a day processing news, as long as I don't have to spend any more time dealing with it. If you force a human to be really careful about dates, or worse, to read the reject logs or argue about software issues in news.software.b, you've wasted more time - human time - the time that matters. mjr.
vnend@phoenix.Princeton.EDU (D. W. James) (05/21/91)
In article <EMV.91May20135904@crane.aa.ox.com> emv@msen.com (Ed Vielmetti) writes: )In article <1991May20.093018@mccall.com> tp@mccall.com (Terry Poot) writes: ) >The standard is almost useless if it isn't conformed to by everybody. ) >Dropping all non-conforming articles makes it perfectly clear what the ) >problem is -- the RFCs aren't being followed. ) Have you ever heard of a defacto standard? The defacto standard for dates ) in usenet messages has for sometime been that set of dates recognized by ) B news, which has apparently been more generous than the RFC. The C news ) change now drops messages that have always been acceptable to every news ) system on the net. )One of the reasons for no longer accepting and trying to parse mangled )dates is that the date parser needed to cope with the multitude of )slightly off date formats is complicated and slow. C News has much )faster date munching routines, but they don't understand every )possible conceivable date that there is. )In this case, the tradeoff is between )passing old news (bad for everyone) vs. dropping malformed articles )has software that puts out mangled dates will be properly coercing )into fixing them, and the net as a whole will be more happy. )--Ed ~sigh~ Does anyone remember "Be generous in what you get, and exact in what you send?" -- Vnend, Lottery winner #316 Ignorance is the mother of adventure. vnend@princeton.edu, vnend@pucc.bitnet, or {backbone}!princeton!nudity!vnend Anonymous posting service (NO FLAMES!):vnend@ms.uky.edu "Q appears, turns the Borg ship into a daytime soap, and cancels it." Frank
billd@fps.com (Bill Davidson) (05/21/91)
In article <1991May20.204313.25340@decuac.dec.com> mjr@hussar.dco.dec.com (Marcus J. Ranum) writes: >I personally don't care if my server spends 5% more time a day processing >news, as long as I don't have to spend any more time dealing with it. Some of us do. Some of us are lucky to get a computer that can handle the load even under ideal software conditions. >If you force a human to be really careful about dates, or worse, to read the >reject logs or argue about software issues in news.software.b, you've >wasted more time - human time - the time that matters. Ah, but most USENET headers are not human generated, including "Date: ". The software generating the article in the first place should have done it right instead of wasting the CPU cycles of hundreds of thousands of machines around the world receiving USENET. For the human generated headers, the posting software should worry about problem and decide wether to drop or correct them. The transport software is the wrong place for it. If the posting software deals with it then it is dealt with once, rather than ? X 10^6. The system I run news on has to do a lot of other things besides news including a lot of interactive stuff. When news is hogging CPU cycles, guess what? The people doing interactive work have to wait longer for responses. There goes the human time. --Bill Davidson
rmk@rmkhome.UUCP (Rick Kelly) (05/21/91)
In article <1991May19.022322.29056@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes: >In article <1991May19.013616.10567@ms.uky.edu> kherron@ms.uky.edu (Kenneth Herron) writes: >>... The message IDs are logged >>along with the complaint , but the messages themselves are thrown away >>(of course) so the concerned administrator has to do a bit of detective work. > >This is something we'd kind of like to do *something* about -- it's not >nice to throw away the evidence -- but defining exactly what should be done >is a little tricky and we haven't yet sorted it all out. How about a news directory called /trashed? The admin could easily go through this and see just what is happening. The down side would be the horrendous problem that causes /usr/spool/news/trashed to overflow the news spool. But hopefully the news admin would have realised there was a problem before it go to this state. Rick Kelly rmk@rmkhome.UUCP frog!rmkhome!rmk rmk@frog.UUCP
) (05/21/91)
billd@fps.com (Bill Davidson) writes: > Ah, but most USENET headers are not human generated, including "Date: ". C News will still dump articles because human-generated headers are wrong, will it not? > The software generating the article in the first place should have done > it right instead of wasting the CPU cycles of hundreds of thousands > of machines around the world receiving USENET. I don't think anyone's contesting that. What we're saying is that if the software generating the article is faulty, someone at that site should be told so that he can fix it, or if the error is trivial it should be corrected and logged locally. My news system here at Mantis writes me little messages saying: 10-Apr-91 05:37 < rnews > | ADDED DATE, Wed, 10 Apr 91 05:37:34 GMT 10-Apr-91 05:38 < rnews > | ADDED DATE, Wed, 10 Apr 91 05:38:02 GMT 10-Apr-91 05:38 < rnews > | ADDED DATE, Wed, 10 Apr 91 05:38:36 GMT 10-Apr-91 05:38 < rnews > | ADDED DATE, Wed, 10 Apr 91 05:38:43 GMT 10-Apr-91 05:38 < rnews > | ADDED DATE, Wed, 10 Apr 91 05:38:49 GMT 10-Apr-91 05:38 < rnews > | ADDED DATE, Wed, 10 Apr 91 05:38:56 GMT It's an MS-DOS BBS system. If it can handle correcting Date: lines, poor little C News ought to be able to. > For the human generated headers, the posting software should worry > about problem and decide wether to drop or correct them. Or perhaps TELL THE USER? mathew
henry@zoo.toronto.edu (Henry Spencer) (05/21/91)
In article <1991May19.102741.24411@wimsey.bc.ca> jhenders@wimsey.bc.ca (John Henders) writes: >... If the first C-news site that rejected >the article dropped it on the floor, this problem wouldn't exist. I think an >attempt to notify the sender is a good idea... There may not be one "first C News site" to see the article; there may be dozens or even hundreds of them, if the article first passed through a B News site with a large news fanout. Once and for all: under no circumstances, none whatsoever, is C News ever going to report problems to a remote site via mail. There is simply no way of making sure that more-or-less innocent people won't get inundated with immense volumes of costly mail. People who make suggestions along those lines can forget it; they have not thought the problem through; we have. We aren't going to do it. The subject is closed. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
scs@iti.org (Steve Simmons) (05/22/91)
henry@zoo.toronto.edu (Henry Spencer) writes: >In article <1991May19.102741.24411@wimsey.bc.ca> jhenders@wimsey.bc.ca (John Henders) writes: >>... If the first C-news site that rejected >>the article dropped it on the floor, this problem wouldn't exist. I think an >>attempt to notify the sender is a good idea... >There may not be one "first C News site" to see the article; there may be >dozens or even hundreds of them, if the article first passed through a >B News site with a large news fanout. Given the multiple connectivity within the network, it is almost guaranteed that *every* c-news site will be offered the defective article, probably once for each non-c-news site which the c-news sites connect to. [[Nota bene -- I'm assuming that rejected articles are not recorded in history, on the off chance that a correct version will come thru from another branch. That'd be the `friendly' thing to do.]] Taking us as an example, that'd be 4 or 5 times. If we assume 40,000 sites (just a guess) with a 25% c-news penetration, thats 10,000 notifications. Goodbye /usr/spool/mail on both the offending system and all its neighbors. Imagine this circumstance. Site foo.bar.baz crunches an article you posted. They corrupt the headers, including the message id. Now 10,000 c-news sites send you mail about it *and you're not the one with the defective software*. The proper response (the *only* proper response) is to notify the sites feeding you bad news that they're doing it. Nothing else. -- "SPAM is a registered trademark of a pork product packed only by Geo. A Hormel & Co. Corp." -- Sun Technical Bulletin, March 1991, pg ii
henry@zoo.toronto.edu (Henry Spencer) (05/22/91)
In article <1991May20.093018@mccall.com> tp@mccall.com (Terry Poot) writes: >Are you saying that this is a problem that doesn't merit attention? ... No, that it's a problem that is very difficult to solve, and that the naive solutions which people constantly propose have unacceptably nasty side effects. >This could have been done with prior announcement, an attempt to find >non-conforming news software and notify users and authors of the impending >change, and some mechanism to ease the transition ... Both announcement and prior investigation were done. There is a limit to how effective the prior investigation could be, however, given very limited manpower and the way B News sites hide problems by rewriting headers. As for transition mechanisms, this is yet another area where things could be improved if we had plenty of paid staff. As it is, we just have to decide "that's good enough -- ship it" at some point. >You could have even done something reasonable, like in this version bitching >about the formatting problem, but accepting it, and in the NEXT version, >rejecting it. That would have given the rest of the net some chance fix our >software BEFORE you started throwing away articles. Our overwhelming experience is that nobody bothers until we start throwing away articles. Warnings simply don't get heeded, not widely and consistently. There is simply too much inertia. >>[preserving evidence] >Maybe you should have thought it out a little further before releasing this >version? See above comments on manpower. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
henry@zoo.toronto.edu (Henry Spencer) (05/22/91)
In article <io49226w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >> You may call it what you wish. Our experience is that all alternatives are >> worse. > >Why, which have you tried? There has been considerable experience, over the years, with news software that sends mail to an article's originator or to a newsgroup's moderator when something untoward is detected. Such software is universally reviled by people who have actually been on the receiving end. >> The losses, while regrettable, are necessary. >Why is it "necessary" to drop articles with the date written as > Date: Mon 20 May 91 15:46 GMT >instead of > Date: Mon, 20 May 91 15:46 GMT >then? Enquiring minds want to know. Because general repair of botched headers is an impossibly hard problem, and imperfect solutions are often cures worse than the disease. This case would not have been hard to deal with... but there is *always* just one more case that would be easy to deal with, and you have to draw the line somewhere. >> You needn't bother repeating them, and I'm not going to bother >> telling you yet again what's wrong with them.) > ... >I hate to inject a note of personal criticism, given that your fan club is >already sending me hate mail, but isn't that just a little bit smug and >self-satisfied? Try tired, short of time, and sick of telling you again and again why your naive ideas don't work or have unacceptable side effects. See my other recent postings for details. Actually, one of your ideas -- using news rather than mail to report the problems -- is not totally naive. It has its own problems and we haven't yet found a variant that is fully workable. (Trying to get everyone to carry a new newsgroup is infeasible, and "control" is a fake newsgroup with non-intuitive semantics.) If we ever find a way to do this, it will probably be along those lines. But if you have thought of it in five minutes, I can just about guarantee that we thought of it long ago. Please *think* about your ideas, including how they might fail, before posting them. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
jbrown@locus.com (Jordan Brown) (05/22/91)
In article <w729224w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >And I have pointed out that certainly in this specific case (my articles >being thrown away due to a bad Date: header) it is trivially easy to fix the >header. Mail systems do it. > [...] >If the error is non-critical and you can still pass on the article, why not >pass it on? The Date: header is _usually_ non-critical - it's _unusually_ for human information only. However, it _is_ one of the fallback mechanisms to deal with problems when broken systems and/or software release garbage onto the net - B news discards articles over a month old to try to squelch floods of old news which might, depending on the failure mode causing them, lead to infinite loops. As such, it's important that it _not_ be fixed incorrectly, because the "fix" could make the problem worse.
igb@fulcrum.bt.co.uk (Ian G Batten) (05/22/91)
In article <wZ49228w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: > Unfortunately, it seems there are quite a few sites out there whose system > administrators have better things to do with their time than read the log > files and send mail messages to people whose articles have been junked. I can say with no shame that I do, in fact, have better things to do with my time than run a news system at all. I suspect the set of people that would prefer the situation ``I read and administer news'' to ``I read news'' is vanishingly small. When mantis starts running a few feeds, and has people phoning up to complain the modems are down or you've dropped their favourite group or somesuch, let us know how it feels. ian
igb@fulcrum.bt.co.uk (Ian G Batten) (05/22/91)
In article <w729224w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: > header. Mail systems do it. Good ones don't. > Deluged with one message, assuming that you don't both forward and report. > Or maybe ten or so messages if you're in a bit of the net which is > exceptionally well-connected via non-C-news sites. There is probably at least a two order of magnitude error in that figure. ian
) (05/22/91)
henry@zoo.toronto.edu (Henry Spencer) writes: > Once and for all: under no circumstances, none whatsoever, is C News ever > going to report problems to a remote site via mail. [...] > We aren't going to do it. The subject is closed. Clearly, then, we have two alternatives: 1. Report problems via some method other than each C News site sending mail; or 2. Try to avoid problems. C News does neither. It therefore needs fixing. mathew
) (05/22/91)
henry@zoo.toronto.edu (Henry Spencer) writes: > In article <1991May20.093018@mccall.com> tp@mccall.com (Terry Poot) writes: > >Are you saying that this is a problem that doesn't merit attention? ... > > No, that it's a problem that is very difficult to solve, and that the > naive solutions which people constantly propose have unacceptably nasty > side effects. I see. And having hours of my time wasted by your software is not an "unacceptably nasty side effect"? > Our overwhelming experience is that nobody bothers until we start throwing > away articles. What experience is this? When did you try mailing out warnings to sites producing bad headers? Or are you just pontificating? mathew
) (05/22/91)
henry@zoo.toronto.edu (Henry Spencer) writes: > In article <io49226w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) > > > The losses, while regrettable, are necessary. > >Why is it "necessary" to drop articles with the date written as > > Date: Mon 20 May 91 15:46 GMT > >instead of > > Date: Mon, 20 May 91 15:46 GMT > >then? Enquiring minds want to know. > > Because general repair of botched headers is an impossibly hard problem, > and imperfect solutions are often cures worse than the disease. This > case would not have been hard to deal with... but there is *always* just > one more case that would be easy to deal with, and you have to draw the > line somewhere. Well, if you don't need to repair the header, don't bother. Just propogate the article. If some other piece of software does need to repair the header, it can do so itself. If you do need to repair the header, do your best. Message-ID: is unsalvageable, Path: is tricky, Newsgroups: is relatively easy, Date: is easy, and you don't need to know about anything else. So there's an excellent case for throwing away articles with malformed message-IDs, and a borderline case for throwing away articles with malformed paths. And that's all. > Actually, one of your ideas -- using news rather than mail to report the > problems -- is not totally naive. It has its own problems and we haven't > yet found a variant that is fully workable. (Trying to get everyone to > carry a new newsgroup is infeasible Why? Tell 'em if they don't carry it you'll drop their articles. That's your usual way to make people conform, isn't it? > and "control" is a fake newsgroup > with non-intuitive semantics.) Do you mean that there is some problem with getting control messages propagated? You could always make the C News reports look like cancel messages for nonexistent articles, or fool sites into propagating them in some other similar way. > Please *think* about your ideas, including how they might fail, before > posting them. Please *think* about your ideas, including how objectionable they might be, before foisting them off on Usenet. mathew
) (05/23/91)
igb@fulcrum.bt.co.uk (Ian G Batten) writes: > In article <wZ49228w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) > > Unfortunately, it seems there are quite a few sites out there whose system > > administrators have better things to do with their time than read the log > > files and send mail messages to people whose articles have been junked. > > When mantis starts running a few > feeds, and has people phoning up to complain the modems are down or > you've dropped their favourite group or somesuch, let us know how it > feels. As a matter of fact, for a while we were providing quite a bit of the UK with its main feed of the UKC-censored newsgroups, and we did get people asking what happened on the occasions when something went wrong. We also get people complaining when the modems are down, since we give dial-in access to selected people. Furthemore, part of our work involves shunting megabytes of data around on a regular basis, and providing support for a large software package which includes an e-mail system. If you think that I am not used to having people phone up and complain, or not used to fighting a losing battle to keep modems and dial-in machines working 24 hours a day, I'm afraid you are mistaken. mathew
) (05/23/91)
igb@fulcrum.bt.co.uk (Ian G Batten) writes: > In article <w729224w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) > > header. Mail systems do it. > > Good ones don't. What do good mail systems do with a bad "Date" header, then? Do they throw away the entire message and not tell anyone at the sender's site? mathew
jmm@eci386.uucp (John Macdonald) (05/23/91)
In article <ben.674596383@dragonfly.wri.com> ben@wri.com (Ben Cox) writes: | |This is the lazy man's answer. Given sufficient effort, one could probably |write a shell script that would fix a Date: header. A Path: header would be |harder, but most others are cake. In a word, bullshit! Given sufficient effort, one could write a shell script that would fix a dozen different bad Date: header formats. There would still be people bitching that *their* invalid date format wasn't being fixed (or that, even worse, it *was* being fixed but the fix was *wrong* and Jan 9 gets converted to Sept 1 by the wrong guess of what 1/9 means). There is no way to write any program that will fix bad format headers and be right every time the fix is *obvious* to the originator and his software. -- sendmail - as easy to operate and as painless as using | John Macdonald manually powered dental tools on yourself - John R. MacMillan | jmm@eci386
ghelmer@dsuvax.uucp (Guy Helmer) (05/24/91)
In <w729224w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >ghelmer@dsuvax.uucp (Guy Helmer) writes: >> In <5ukZ24w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >> >chip@tct.com (Chip Salzenberg) writes: >> >> True, C News drops articles on the floor. But it logs having done so, >> >> and it notifies the Usenet administrator in the output of newsdaily. >> >> >So what the hell use is that? None of those Usenet administrators told *ME* >> >about it, and I'm the one who has to fix my software. >[ Point not answered ] I don't have the message references handy, but I've seen postings from at least one admin that said he sent you mail about your non-conformant news headers when his machine dropped them. >> >If an article can be corrected to conform, then it should be, and should the >> >be delivered. >> >> Henry has pointed out many times that it is difficult to correct >> problems in an article. >And I have pointed out that certainly in this specific case (my articles >being thrown away due to a bad Date: header) it is trivially easy to fix the >header. Mail systems do it. So fix your own header before any conforming software downstream has to look at it. >> No. You better not pass on the article, or you're part of the problem. >If the error is non-critical and you can still pass on the article, why not >pass it on? Software which falls over in a heap as a result of bad input is >extremely broken, and C News should not attempt to Nanny it. C-news doesn't fall over when it sees bad input, it dumps the input. Software which ignores a bad input is not broken. >> Attempts to notify the article originator would be a bad thing to do, too. >> A site producing news may not receive mail or there may not be a path >> for a message to follow back to the sender. >In which case the notification will fail to get through. Fair enough. Mail >bounce messages sometimes fail to get through, but that doesn't mean you >shouldn't try to send them. Worst case senario: a machine suddenly munges the headers of a huge batch of old news and dumps it into a backbone site that isn't dropping improperly formed articles. The backbone site feeds 20 or sites through full feeds, which each feed another 10 to 20 sites each. Eventually, between twenty and a few thousand sites running your "improved" software will notice these articles and each site will think it's doing the "right thing" by trying to notify the original authors. At this point, everyone could sit back and watch the net eat itself. If the net ever recovered, the original authors of the old news would be drowning under mail from stupid news systems incorrectly pointing the finger at these poor authors. This senario isn't far fetched. Think about how many times over the past year sites have munged headers and injected bad articles into the news stream. >> Worse yet, if you attempt to both pass on the article and notify the sender, >> you and everyone running software like yours will clog the net with tons, er, >> megabytes, of useless messages. >Fair enough. So either pass on the article or notify the sender. I'm happy >with either. All I'm not happy with is not passing on the article AND not >notifying the sender. Again, the way I heard it, you were notified. >> And what about a message that's had it's headers munged by a single machine >> running junky software? A person would be deluged with messages from >> stupid news software incorrectly attributing the bad article to him/her >> and his/her software. >Deluged with one message, assuming that you don't both forward and report. >Or maybe ten or so messages if you're in a bit of the net which is >exceptionally well-connected via non-C-news sites. Refer back to the "worst case senario". You cannot back up a claim that only "ten or so messages" would be sent. >> >Rubbish. Throwing away data containing correctable errors, without issuing >> >any warning message, is NEVER the right thing. >> >> There are warning messages. Be a good admin---correctly configure your >> machine's news and mail software (i.e., have news send mail about problems >> to usenet and alias usenet to the news admin id, but don't depend on news >> to tell you when it is having trouble)---and watch your system. >I do. I read our system logs, I check for mail bounces and assist, I respond >to queries, I check out errors. Great! It sounds to me like your downstream sysadmin also watches his system, and tried to inform you when your system was producing non-conformant articles. That's what I meant by "be a good admin". >Unfortunately, there isn't a log file on our system which will tell me >whether UUNET's copy of C News is throwing away all my articles. If there >were, there would be no problem. >ALL I AM ASKING is that if a posting from a site X is dropped, SOMEONE from >site X should be told about it. This notification should NOT rely on some >random system administrator happening to take pity on site X. >> Dump your self-righteous attack against C-news. News isn't mission-critical; >> news isn't life-sustaining; delivery of articles isn't guaranteed. >True. However, delivery of articles is much less guaranteed now that C News >works in "drop all data on error" mode. >As to news not being important -- some people spend literally hours writing >stuff to post to Usenet. I'd rather that they continued to do so. And they >won't if they think that some big C News site will throw it all away because >they made a typo in a header field. >> The RFC's >> exist, and if your software doesn't implement them correctly, that's tough. >Interesting attitude. Am I to take it that you would not object if I wrote a >program to forge cancel messages for all articles which were not STRICTLY RFC >compliant? That actually might not be too bad of an idea. :-) It would get a lot of people's attention very quickly. >> Check your software to make sure it generates RFC-compliant articles. That's >> the only way you can be guaranteed that you have a good chance of getting >> your article distributed throughout the net. >So what if I'm an arts graduate? Do you really expect every random user of >Usenet to read through the RFC documents and check the news software which >exists on his system? I'd expect a news admin to be aware of whether his software has problems. With the finite number of news packages available, news admins can quickly become aware of problems (especially like this "bad header" discussion) by just reading the news. >> You can't blame anyone or >> any software for dumping your article and not telling you about it, >> especially if your software is faulty. >Of course I can. C News did not need to dump my article. It deliberately >chose to do so, however. It also deliberately chose not to warn me, and so >the error continued. Therefore it was at least partly C News' fault that the >articles were lost. You are entitled to your opinion. I don't agree. >mathew -- Guy Helmer, Dakota State University Computing Services helmer@sdnet.bitnet, dsuvax!ghelmer@wunoc.wustl.edu, ghelmer@dsuvax.dsu.edu "Everybody need a soft filter / Everybody need reverse polarity" - Rush
jbuck@forney.berkeley.edu (Joe Buck) (05/24/91)
In article <9105202307.10@rmkhome.UUCP>, rmk@rmkhome.UUCP (Rick Kelly) writes: |> How about a news directory called /trashed? The admin could easily go through |> this and see just what is happening. The down side would be the horrendous |> problem that causes /usr/spool/news/trashed to overflow the news spool. But |> hopefully the news admin would have realised there was a problem before it |> go to this state. Rick, you're reinventing the wheel. There's already a news directory called /usr/spool/news/junk that serves the purpose. Why not use what's already there? -- Joe Buck jbuck@galileo.berkeley.edu {uunet,ucbvax}!galileo.berkeley.edu!jbuck
rickert@mp.cs.niu.edu (Neil Rickert) (05/24/91)
In article <1991May24.014500.12255@agate.berkeley.edu> jbuck@forney.berkeley.edu (Joe Buck) writes: >In article <9105202307.10@rmkhome.UUCP>, rmk@rmkhome.UUCP (Rick Kelly) writes: >|> How about a news directory called /trashed? The admin could easily go through > >Rick, you're reinventing the wheel. There's already a news directory called >/usr/spool/news/junk that serves the purpose. Why not use what's already >there? This misses the point. It really doesn't matter whether you put these articles in junk, or in trashed, or in the appropriate directory for the news group. What matters is that they never be forwarded to any additional sites. The one benefit of a separate directory (trashed) would be that you could put the articles their, but not enter them in the history file. That way if a valid copy arrived later you would still see it. You would probably need to have 'newsdaily' discard all old trashed articles. You might even want 'spacefor' to do so when space gets tight. ersonally, I am not convinced this is a good idea - I see several problems. But it is not the same as junk. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
tim@dell.co.uk (Tim Wright) (05/24/91)
In <wZ49228w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >chip@tct.com (Chip Salzenberg) writes: >> Administrators who get "bogus header" messages and don't pass them on >> deserve flamage. >Unfortunately, it seems there are quite a few sites out there whose system >administrators have better things to do with their time than read the log >files and send mail messages to people whose articles have been junked. You don't need to read the log files. I take it you do have access to awk, cron and mail. It can't be that hard to put together a small summarising script to mail you, can it ?? Quite a few sites conversely do take the time. >As a general principle, it is unwise to design software in such a way that it >relies on system administrators being conscientious individuals. Perhaps you'd care to mention that to Ritchie, Thompson et.al. some time. (1/2 :-) One of the biggest obstacles to improving usenet over time has been the problem of shoddy software. Whilst it is regrettable that some sites lost some articles, it was quite amazing to see how quickly the volume of bad articles dropped. Do you really expect the shell to magically rescue you if you type 'rm -rf *' as root ? Why then should the news transport try to fix up your mistakes. As has been pointed out, there is no reliable way to do this which may not break other conforming software. If you really spend hours preparing articles, surely you can spend 30 seconds verifying the validity of your headers. Novice users should not go around editing headers anyway unless they have read the specs and deserve to have their articles dropped if they don't. The only sympathy I can have is if someone is using common and supposedly correct news software and finds that it is generating bad headers. I do not believe that there have been many/any examples of this. I would welcome counterexamples. Tim -- Tim Wright, Dell Computer Corp., Bracknell | Domain: tim@dell.co.uk Berkshire, UK, RG12 1RW. Tel: +44-344-860456 | Uucp: ...!ukc!delluk!tim Smoke me a Kipper, I'll be back for breakfast - Red Dwarf
tanner@cdis-1.compu.com (Dr. T. Andrews) (05/24/91)
On having a volunteer project instead of paid staff for C news...
hs-) As it is, we just have to decide "that's good enough -- ship it"
hs-) at some point.
The same applies in commercial products, by the way. Don't feel
that commercial packaging means that first we get everything
perfect, and then we ship.
We're about to ship BJ 5.0; if you think that there won't be a
5.1, it may be because you don't know that we've already started
planning for it.
Let me also tell you that the customers will be grateful when
they get 5.0; it will be faster (in some cases much faster), and
use less disk space for the database, and cost less to produce.
More features will be included in response to customer requests.
They'll also be glad to get 5.1, when that comes out. Still,
they don't want to wait until we do all of the scheduled 5.1 work
before they get 5.0; they want us to say ``That's good enough --
ship it'' and they want us to say it fairly soon.
--
uflorida!ki4pv!cdis-1!tanner {uunet dsinc}!cdin-1!cdis-1!tannerhenry@zoo.toronto.edu (Henry Spencer) (05/24/91)
In article <FDmc311w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >> No, that it's a problem that is very difficult to solve, and that the >> naive solutions which people constantly propose have unacceptably nasty >> side effects. > >I see. And having hours of my time wasted by your software is not an >"unacceptably nasty side effect"? It's regrettable, but not in the same league. >> Our overwhelming experience is that nobody bothers until we start throwing >> away articles. > >What experience is this? When did you try mailing out warnings to sites >producing bad headers? Or are you just pontificating? Ever looked at the statistics on who runs what? There are still sites out there running versions of news so old they creak, versions which are known to have nasty bugs in them. On a lesser scale, it's not uncommon for us to get problem reports that we recognize as caused by things we fixed many months ago. The inertia in the net's software is enormous; this is a long-established fact. (And yes, as a matter of fact we did try some mailings before releasing the March patches. They had *some* effect.) -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
henry@zoo.toronto.edu (Henry Spencer) (05/24/91)
In article <57Lc310w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >Clearly, then, we have two alternatives: >1. Report problems via some method other than each C News site sending mail; >or >2. Try to avoid problems. We do both. Neither is done perfectly. Improvements are expected. (By the way, one very good aid to avoiding problems is to prevent them from spreading.) -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
chip@tct.com (Chip Salzenberg) (05/25/91)
According to mathew@mantis.co.uk (CNEWS MUST DIE!): >Well, if you don't need to repair the header, don't bother. Just propogate >the article. If some other piece of software does need to repair the header, >it can do so itself. You don't realize the importance of the Date header. Older C News versions did as you suggest. As a result, they were vulnerable to the loads of old news which get dumped on the net periodically due to admin inexperience and mistakes. C News insists on a parseable Date so it can guarantee that the news it forwards isn't old junk. >Please *think* about your ideas, including how objectionable they might be, >before foisting them off on Usenet. Get a grip. Geoff and Henry don't force administrators at gunpoint to apply C News patches. AAdministrators who apply C News patches do so because they decide to do so. If you don't like the way *my* site works, take it up with *me*. Leave them out of it. -- Chip Salzenberg at Teltronics/TCT <chip@tct.com>, <uunet!pdn!tct!chip> perl -e 'sub do { print "extinct!\n"; } do do()'
billd@fps.com (Bill Davidson) (05/26/91)
In article <57Lc310w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >henry@zoo.toronto.edu (Henry Spencer) writes: >> Once and for all: under no circumstances, none whatsoever, is C News ever >> going to report problems to a remote site via mail. [...] >> We aren't going to do it. The subject is closed. >Clearly, then, we have two alternatives: >1. Report problems via some method other than each C News site sending mail; >or >2. Try to avoid problems. >C News does neither. It therefore needs fixing. Actually, it does (1). It reposrts the problem in a log. It's not broken. It doesn't need fixing. Other posting software is broken. It does need fixing. --Bill Davidson
tp@mccall.com (Terry Poot) (05/26/91)
In article <tim.675083093@holly>, tim@dell.co.uk (Tim Wright) writes: >In <wZ49228w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >>chip@tct.com (Chip Salzenberg) writes: >>> Administrators who get "bogus header" messages and don't pass them on >>> deserve flamage. > >>Unfortunately, it seems there are quite a few sites out there whose system >>administrators have better things to do with their time than read the log >>files and send mail messages to people whose articles have been junked. > >You don't need to read the log files. I take it you do have access to awk, >cron and mail. It can't be that hard to put together a small summarising >script to mail you, can it ?? Quite a few sites conversely do take the time. In my case, 2. If you have such a tool, why don't you send it to Henry or Geoff so they can include it with C news? C news does not include such a facility. Do you actually believe even a majority of C news admins will take the time to write one? >One of the biggest obstacles to improving usenet over time has been the >problem of shoddy software. Whilst it is regrettable that some sites lost >some articles, it was quite amazing to see how quickly the volume of bad >articles dropped. You want traffic volume to go down? Do a parity check on the message ID, and drop the article if the parity is even. It will have amazing results on bandwidth. Of course, that, too, is antisocial. >If you really spend >hours preparing articles, surely you can spend 30 seconds verifying the >validity of your headers. Novice users should not go around editing headers >anyway unless they have read the specs and deserve to have their articles >dropped if they don't. Are you saying you do or don't want novice users to attempt to correct the headers in ways C news likes? I've been talking about headers that are valid to B news, but not C news. You seem to be saying that the user should verify his headers, but that he should be required to read and understand the RFC's before doing so. Are all your users required to read RFC822 and RFC1036 before using news? Mine aren't, and I don't think they should be. >The only sympathy I can have is if someone is using common and supposedly >correct news software and finds that it is generating bad headers. I do not >believe that there have been many/any examples of this. I would welcome >counterexamples. I am a user (not author) of the mxrn newsreader, which generated B news format dates that C news did not like. I found out about this whole issue this last week, and have generated a patch. I've sent the patch out and hopefully in the near future, other sites running mxrn (or dxrn, the other variant built from the same sources) will have the patch installed. In the mean time, all C news sites are dropping all messages from users of this software, and I'd venture to guess that almost none of them know about it, as nobody has made any effort to find them and tell them about it. I've attempted to post the patch widely, but I imagine that this problem will persist for quite some time, given the general problem of trying to get software updated on the net. If you think I should have known about this before, look again at the subject line of this thread. Exactly 2 news admins, world wide, told me there was a problem. Thus, obviously, logging these things as a method of reporting has failed miserably. (I have since gotten mail from at least 2 news admins that knew I had a problem, and DIDN'T tell me.) Given the wide range of date formats that B news handles, did you honestly believe that there is no news software that generates any of them but the one in RFC822? In article <tim.675083877@holly>, tim@dell.co.uk (Tim Wright) writes: >In <1991May20.141852.30919@kuhub.cc.ukans.edu> >sloane@kuhub.cc.ukans.edu writes: >>In what groups were these warnings posted? I don't recall seeing >>them. If the only group was news.software.b, then I can understand >>why poeple not runing b or c news wouldn't see them. Or are you >>contending that everyone should read n.s.b just in case some change >>was made that would cause them problems? Do you read n.s.anu-news? > >No, but since the majority of the net is made up of B and C news sites >and news.software.b carries articles of importance to all net users I >would say that any VMS news sysadmin not reading it was foolish. What an incredibly chauvinistic attitude! The whole point of separate newsgroups is so people that aren't concerned with a topic don't have to read it. If everyone needed to know everything about B news and C news, it'd be discussed in news.admin. On the contrary, when B news and C news people are going to do something that affects other than their own user community, it is appropriate for them to seek to reach that other community, not the other way around. In this case, that means news.admin as a minimum. In this particular case, I happen to think that news.announce.important would have been in order. (That is, theoretically, the only group that supposedly all usenet sites carry, and probably has an extrememly high readership, possibly enough to have avoided most of the problems, and maybe even a lot of this debate.) I can quite readily guarantee that if an announcement of this had been posted there well before this action were taken, I'd (or the author) would have fixed dxrn/mxrn well before the C news patches were installed. BTW, I think a posting to news.announce.important is still in order. It won't avert problems, as an early posting would have done, but it may help people who don't know their messages are being dropped find out about it and take some appropriate action. If you believe these people will be told by some kindly C news admin that their articles are being rejected, you are hopelessly naive. BTW, I applaud the list that scott@zorch posted to news.lists. However, I suggest it be retitled. The subject line looks like a C news bug report, and in any case totally uninteresting to non-C news admins. I only looked at it because it was mentioned here. I would have skipped it. Also, I'm not in it. Since C news dropped my articles (before I patched mxrn) they presumably never made it to zorch. Thus, one such list will only find a small minority of the sites having the problem. If, however, this had been done as I have suggested, by putting in the messages one release before discarding the articles, this would have found many more problems. I'm interested if Henry has considered this, and what flaw it has that I'm overlooking. Lists could have been collected from large sites around the net and merged (by a program of course), producing a reasonable list of sites generating bogus headers. It wouldn't be complete, but would have been better than nothing. (Yes, I know the suggestion was after the fact, but I didn't know about this issue before the fact.) -- Terry Poot <tp@mccall.com> The McCall Pattern Company (uucp: ...!rutgers!ksuvax1!deimos!mccall!tp) 615 McCall Road (800)255-2762, in KS (913)776-4041 Manhattan, KS 66502, USA
rickert@mp.cs.niu.edu (Neil Rickert) (05/26/91)
In article <1991May26.082113@mccall.com> tp@mccall.com (Terry Poot) writes: >the RFC's before doing so. Are all your users required to read RFC822 and >RFC1036 before using news? Mine aren't, and I don't think they should be. I don't think users should be required to read RFC1036 either. But I do think news software developers should, and should make sure that articles entered into the news system with their software are compliant. >Given the wide range of date formats that B news handles, did you honestly >believe that there is no news software that generates any of them but the >one in RFC822? You are missing the point. There is a wide range of date formats that C news handles too. And when an article is submitted in one of these formats it is reformatted into an RFC compliant form. It just so happens that the design of B-news is such that this reformatting occurs at every B-news site, not just on article submission. If somebody produces software which incorrectly assumes that it can put out bad dates and rely on other software somewhere else to fix it, you should not blame C-news for the predictable result. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
billd@fps.com (Bill Davidson) (05/27/91)
In article <1991May26.082113@mccall.com> tp@mccall.com (Terry Poot) writes: >Are you saying you do or don't want novice users to attempt to correct the >headers in ways C news likes? I've been talking about headers that are >valid to B news, but not C news. There's a big difference between valid and acceptable. Bnews accepts some types of invalid headers. I believe it tries to fix some of them. >You seem to be saying that the user should >verify his headers, but that he should be required to read and understand >the RFC's before doing so. Are all your users required to read RFC822 and >RFC1036 before using news? Mine aren't, and I don't think they should be. Where do you get this stuff? Nobody said anything like that. Most headers are created by the posting software and any user created headers should be verified as proper and possibly fixed by the posting software, not the transport software. People who write posting software certainly SHOULD read RFC822 and RFC1036. If you're using broken software, you should fix it or change to software that isn't broken. >Given the wide range of date formats that B news handles, did you honestly >believe that there is no news software that generates any of them but the >one in RFC822? Nobody thought that. Bnews handles invalid headers not because it's right to do so but in order to avoid having all the bozos with broken software from complaining. Well, now you are. It would have been better if Bnews had never accepted bad headers. That way the inevitable pain would have been out of the way years ago and involved many fewer sites. At some point it had to happen. No time like the present. --Bill Davidson
scs@iti.org (Steve Simmons) (05/27/91)
tp@mccall.com (Terry Poot) writes: >In article <tim.675083093@holly>, tim@dell.co.uk (Tim Wright) writes: >>In <wZ49228w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >>>chip@tct.com (Chip Salzenberg) writes: >>>> Administrators who get "bogus header" messages and don't pass them on >>>> deserve flamage. >> >>>Unfortunately, it seems there are quite a few sites out there whose system >>>administrators have better things to do with their time than read the log >>>files and send mail messages to people whose articles have been junked. >> >>You don't need to read the log files. I take it you do have access to awk, >>cron and mail. It can't be that hard to put together a small summarising >>script to mail you, can it ?? Quite a few sites conversely do take the time. >In my case, 2. If you have such a tool, why don't you send it to Henry or >Geoff so they can include it with C news? C news does not include such a >facility. Do you actually believe even a majority of C news admins will >take the time to write one? The following script generates a detailed report on news traffic and problems including header reports. It's by no means done, but it seems clear there's a need...the folks who've been beta testing it seem to like it. Patches will be appreciated, but be forewarned it's going to get a major rewrite Real Soon Now. #! /bin/sh # This is a shell archive. Remove anything before this line, then unpack # it by saving it into a file and typing "sh file". To overwrite existing # files, type "sh file -c". You can also feed this as standard input via # unshar, or by typing "sh <file", e.g.. If this archive is complete, you # will see the following message at the end: # "End of shell archive." # Contents: cnews.logrep # Wrapped by scs@wotan.iti.org on Sun May 26 23:59:04 1991 PATH=/bin:/usr/bin:/usr/ucb ; export PATH if test -f 'cnews.logrep' -a "${1}" != "-c" ; then echo shar: Will not clobber existing file \"'cnews.logrep'\" else echo shar: Extracting \"'cnews.logrep'\" \(21554 characters\) sed "s/^X//" >'cnews.logrep' <<'END_OF_FILE' X: /bin/sh X# X# Shell script for cnews log reports X# X# This script scans the cnews data areas and looks for things that X# are out of whack. It reads some of the log files looking for X# interesting information. If there are no errors or oddities, it X# tells you so. This should be run immediately after the cnews daily X# 'newsdaily' script. It expects the logs to be named <thing>.o X# but you can change that by resetting the *LOG variables as shown X# below. X# X# One point of "philosophy": this whole script is built on the idea X# that silence is *not* golden. If things are OK, it explicitly X# tells you so. X# X# The log file analyzer is a big awk script. Make sure you understand X# awk associative arrays before attempting to modify it. X# X# $RCSfile: cnews.logrep,v $ $Revision: 0.16 $ X# X# $Author: news $ $Date: 91/04/22 11:15:10 $ X# X# $State: Exp $ $Locker: news $ X# X# $Log: cnews.logrep,v $ X# Revision 0.16 91/04/22 11:15:10 news X# Cleaned up extra backslashes. Added sendme processing with the ihaves. X# Made names of all log files explicit. X# X# Revision 0.15 91/04/22 10:03:59 news X# Added cancel and failed cancel reporting. X# X# Revision 0.14 91/04/22 09:12:14 news X# Got oldies number right. X# X# Revision 0.13 91/04/12 11:40:12 news X# Added reports on future dates and too-old dates, added notice on X# postings which lack message ids. X# X# Revision 0.12 91/04/11 16:46:08 news X# Added outgoing batch checking. X# X# Revision 0.11 91/04/11 16:33:33 news X# Many changes as per suggestions from Brendan Kehoe and Owen Medd. X# Many many changes. Many many many changes. X# X# Initialize the dirs, names of the logs, and variables X# X. /usr/lib/news/cnewsbin/config XPATH="$NEWSPATH:$NEWSBIN" Xexport PATH XSCRIPT=`basename $0` X# X# Shorthand names for standard C news locations X# XLOGDIR=$NEWSCTL XSYSFILE=${LOGDIR}/sys XINCOMING=${NEWSARTS}/in.coming XOUTGOING=${NEWSARGS}/out.going XBADBATCHES=${INCOMING}/bad X# X# If you are running RELAYNEWS as a daemon, uncomment this next line X# X#RELAYNEWS=1 X# X# When there are completely unrecognizable entries in the log, the X# normal procedure is to print them out in toto so you can extend X# this script to say something intelligent about them. But if X# there are hundreds, it can be overwhelming. These variables allow X# you to restrict the number you will print. X# XMAX_DEFECTS=5 XMAX_UNKNOWN=5 XMAX_MYSTERY_DASH=5 XMAX_MYSTERY=5 XMAX_NOPATH=5 X# X# If this is not where you put the old log files, adjust X# these locations and names accordingly X# XERRLOG=${LOGDIR}/errlog.o XNEWSLOG=${LOGDIR}/log.o XBATCHLOG=${NEWSCTL}/batchlog.o XBATCHPARMS=${NEWSCTL}/batchparms X# XNEWSMGR=news XHOSTNAME=`newshostname` XTEMPFILE=/tmp/${SCRIPT}.$$ Xset `date` XDATE="${1} ${2} ${3}, ${6}" Xtrap 'rm -f $TEMPFILE ; exit' 0 1 2 3 15 Xexec > $TEMPFILE X# X# Print preface. X# Xcat << EOF XThis is the news system status report for $DATE as Xgenerated by $0. X XEOF X# X# Errorcheck to be sure the error log exists. X# Xif [ "" = "$HOSTNAME" ] ; then X cat << EOF XThe news host name was not found. Please check the file X$NEWSCTL/whoami, and make sure the host name is set Xappropriately for your system (e.g, look at \'uuname -l\' and X\'uname -n\'. The report will be somewhat incomplete. X XEOF Xfi X# X# Errorcheck to be sure the error log exists. X# Xif [ ! -f "$ERRLOG" ] ; then X cat << EOF XThe news error log ($ERRLOG) for was not found. XPlease check the news system and this script. X XEOF Xelse X # X # If the error log is empty, report that. Otherwise copy it to X # the news manager. In either case, give a nice prefatory remark. X # X SIZE=`wc -l < $ERRLOG` X SIZE=`echo $SIZE` X if [ $SIZE = 0 ] ; then X cat << EOF XThe news error log ($ERRLOG) shows no errors. X XEOF X else X # X # Give a nice prefatory remark, then copy the log to the news X # manager. Put beginning and end markers. X # X cat << EOF XThe news error log ($ERRLOG) has $SIZE complaints. A Xcopy of the error log is included below: X X$ERRLOG: XEOF X cat < ${ERRLOG} X cat << EOF XEnd of $ERRLOG X XEOF X fi Xfi X# X# Report on bad batches left lying around X# Xif [ -d ${BADBATCHES} ] ; then X ( X cd $BADBATCHES X COUNT=`ls | wc -l` X COUNT=`echo $COUNT` X if [ $COUNT = 0 ] ; then X echo "There are no bad batches being held in $BADBATCHES." X else X echo "There are $COUNT bad batches being held in $BADBATCHES:" X if [ $COUNT -gt 10 ] ; then X ls -C X else X ls -l X fi X fi X ) Xelse X echo "Could not find bad batches directory ($BADBATCHES)!" Xfi Xecho "" X# X# Report on old nrun files. X# Xif [ ! -d $INCOMING ] ; then X echo "Could not find incoming directory ($INCOMING)!" Xelse X ( X cd $INCOMING X COUNT=`find . -name 'n*' -type f -mtime +1 -print | wc -l` X COUNT=`echo $COUNT` X if [ $COUNT != 0 ] ; then X echo "There seem to be old run fragments left in ${INCOMING}:" X echo "" X ls -ls n* X else X echo "There are no old run fragments left in ${INCOMING}." X fi X echo "" X # X # Report on old batch files. X # X LIST=`find . -name "[1-9]*" -type f -mtime +1 -print` X COUNT=`echo $LIST | wc -w` X COUNT=`echo $COUNT` X if [ $COUNT != 0 ] ; then X cat << EOF XThere seem to be $COUNT old incoming batches left in ${INCOMING}. XYou should check to see if unbatching is being done properly. The Xold batches are: X XEOF X if [ $COUNT -gt 10 ] ; then X ls -C $LIST X else X ls -ld $LIST X fi X else X echo "There are no old batches left in ${INCOMING}." X fi X ) Xfi Xecho "" X# X# Report on the status of batching. If the batchparms files X# does not exist, simply state that we think no batching is X# going on. If it's there, try to report on batching. If your X# site doesn't do batching and you'd like to shut up this section X# of the report, rename $BATCHPARMS to $BATCHPARMS.sample. X# X# This needs to be expanded to report more fully on batch results.... X# Xif [ ! -f "$BATCHPARMS" ] ; then X echo "You do not appear to be doing any uucp batching (no batch" X echo "paramters file $BATCHPARMS)." Xelse X if [ ! -f "$BATCHLOG" ] ; then X echo "Could not process batch log ${BATCHLOG}." X else X # X # Report on stalled batching for outside sites. Report if the X # log does not exist. X # X COUNT=`grep "no recent movement" ${BATCHLOG} | wc -l` X COUNT=`echo $COUNT` X if [ $COUNT != 0 ] ; then X echo "The news batch queue is full for the following sites:" X grep "no recent movement" ${BATCHLOG} | awk '{ printf "%s\n", $1 }' | sort | uniq X else X echo "There are no full outgoing batch queues show in $BATCHLOG." X fi X fi Xfi Xecho "" X# X# Errorcheck to be sure the regular log exists. X# Xif [ ! -f "$NEWSLOG" ] ; then X cat << EOF XCould not process news log! X XThe standard news log ($NEWSLOG) was not found. XPlease check the news system and this script. X XEOF Xelse X # X # Generate traffic, activity, and oddity report. I am X # not a big awk fan, but gotta admit you can do a lot X # with associative arrays. X # X awk < $NEWSLOG ' XBEGIN { X accept_count = 0 X entry_count = 0 X ship_count = 0 X x_count = 0 X duplicate_count = 0 X ihave_count = 0 X sendme_count = 0 X unapproved_count = 0 X unsub_count = 0 X junk_count = 0 X local_articles = 0 X bad_header_count = 0 X empty_header_count = 0 X non_header_count = 0 X bad_date_count = 0 X no_msgid_count = 0 X unknown_no_count = 0 X future_count = 0 X oldies_count = 0 X cancel_count = 0 X precancel_count = 0 X nopath_count = 0 X mystery_dash_count = 0 X mystery_count = 0 X defect_count = 0 X hostname = "'$HOSTNAME'" X} X{ X # X # Count all entries, note and save defective ones X # X entry_count++ X if ( NF < 6 ) X { X defect_count++ X defective_entry[ $0 ]++ X } X else if ( $5 == "+" ) X { X # X # Track the accepted articles X # X accept_count++ X accept_host[ $4 ]++ X if ( NF > 6 ) X { X ship_count++ X shipfield = 7 + "'$RELAYNEWS'" X # X # Record which systems got articles X # X while ( shipfield <= NF ) X { X ship_list[ $shipfield ]++ X shipfield++ X } X } X # X # Track the number of local postings X # X if ( $4 == hostname ) X local_count++ X } X # X # Track all junked articles by reason junked. X # X else if ( $5 == "j" ) X { X junk_count++ X reason = 7 X junkfor="" X while ( reason <= NF ) X { X junkfor = junkfor $reason " " X reason++ X } X junk_reason[ junkfor ]++ X } X # X # Track ihave/sendme records X # X else if ( $5 == "i" ) X { X ihave_count++ X ihave_source[ $4 ]++ X } X else if ( $5 == "s" ) X { X sendme_count++ X sendme_source[ $4 ]++ X } X # X # Track the exceptions on a class-by-class basis X # X else if ( $5 == "-" ) X { X # X # Track the number of duplicates we get from various sites. X # X if ( $7 == "duplicate" ) X { X duplicate_count++ X duplicate_host[ $4 ]++ X } X # X # Track the number of unapproved articles and their sources. X # X else if ( $7 == "unapproved" ) X { X unapproved_count++ X unapproved_source[ $4 ]++ X unapproved_target[ $12 ]++ X } X # X # Lots of errors are no this, no that, etc. We process X # them all at once. X # X else if ( $7 == "no" ) X { X if ( $8 == "subscribed" ) X { X # X # Track the number of unsubscribed articles we X # get from various sites and the target group. X # X unsub_count++ X unsub_source[ $4 ]++ X unsub_target[ $11 ]++ X } X # X # These are needed for new pedantic Cnews. Over X # time these should decline. We track where the X # articles came from but not the message ids. If X # the admin is really interested, they can track X # them down by grepping the logfile for "no FOO" X # and the name of the offending host. X # X else if ( ( $8 == "Date:" ) || ( $8 == "@" ) || ( $8 == "From:" ) || ( $8 == "Subject:" ) ) X { X bad_header_count++ X bad_header_source[ $4 ]++ X } X else X # X # These are minus for unknown reason. An obvious X # candidate for mods to this script. X # X { X unknown_no_count++ X unknown_no_field[ $8 ]++ X unknown_no_entry[ $0 ]++ X } X } X # X # This next is a *severe* error! X # X else if ( ( $6 == "no" ) && ( $7 == "Message-ID:" ) ) X { X no_msgid_count++ X no_msgid_source[ $4 ]++ X } X # X # Other less serious errors X # X else if ( $7 == "empty" ) X { X empty_header_count++ X empty_header_field[ $8 ]++ X empty_header_source[ $4 ]++ X } X else if ( ( $9 == "contains" ) && ( $10 == "non-header" ) ) X { X non_header_count++ X non_header_source[ $4 ]++ X } X else if ( ( $7 == "unparsable" ) && ( $8 == "Date:" ) ) X { X bad_date_count++ X bad_date_source[ $4 ]++ X } X # X # Track the number of articles which we explicitly reject via X # x records in the active file. We do not track the specific X # groups for two reasons -- presumably since you xed it, you X # know what you reject; and it is hard to parse. X # X else if ( ( $7 == "all" ) && ( $8 == "groups" ) ) X { X x_count++ X } X # X # Track the articles which are too far in the future X # X else if ( ( $7 == "Date:" ) && ( $8 == "too" ) && ( $9 == "far" ) ) X { X future_count++ X future_source[ $4 ]++ X } X # X # Track the articles which are too old to be worth keeping X # I love this cnews feature -- it will probably keep things X # freer of loops than anything else. X # X else if ( ( $7 == "older" ) && ( $8 == "than" ) ) X { X oldies_count++ X oldies_source[ $4 ]++ X } X # X # Some articles come in sans Path: headers. This X # identifies them. X # X else if ( ( $7 == "no" ) && ( $7 == "Path:" ) && ( $8 == "header" ) ) X { X nopath_count++ X nopath_source[ $4 ]++ X } X # X # Any unrecognised "-" tag is kept here. As we find these X # they should be added to the things handled above. X # X else X { X mystery_dash_count++ X mystery_dash_entry[ $0 ]++ X } X } X # X # This counts cancel messages. They are not generated by X # standard C news, but by Dave Aldens relaynews daemon. X # X else if ( $5 == "c" ) X { X cancel_count++ X } X # X # This counts failed cancel messages. They are not generated X # by standard C news, but by Dave Aldens relaynews daemon. X # X else if ( $5 == "f" ) X { X precancel_count++ X } X # X # Any unrecognised tags get noted here. As we get these X # they sould be added to the things handled above. X # X else X { X mystery_count++ X mystery_tags[ $5 ]++ X mystery_entry[ $0 ]++ X } X} XEND { X printf "\nThere were %d entries in the standard log. Breakdown:\n", entry_count X if ( entry_count != 0 ) X { X # X # Report the data by categories. X # X # Local postings. It would be nice to report the newsgroups X # posted to, but that data is not in the log. X # X printf "\n%6d articles were posted from this site (%s)\n", local_count, hostname X # X # Next, where outside articles came from X # X printf "\n%6d incoming articles accepted for processing\n", accept_count X if ( accept_count > 0 ) X for ( host in accept_host ) X printf " %6d from %s\n", accept_host[ host ], host X printf "\n%6d of those were rejected as duplicates\n", duplicate_count X # X # Report number of duplicates and who gave them to us. X # X if ( duplicate_count > 0 ) X for ( host in duplicate_host ) X printf " %6d from %s\n", duplicate_host[ host ], host X # X # Ihave activity. X # X printf "\n%6d/%d ihave/sendme messages were processed\n", ihave_count, sendme_count X if ( ihave_count > 1 ) X { X for ( source in ihave_source ) X printf " %6d from %s\n", ihave_source[ source ], source X } X if ( sendme_count > 1 ) X { X for ( source in sendme_source ) X printf " %6d from %s\n", sendme_source[ source ], source X } X # X # Cancel and failed cancel reporting. These messages only appear X # in the log file if you are running Dave Aldens relaynews X # daemon. Since no cancel messages almost certianly means X # vanilla C news, we do not report on zero counts. X # X if ( cancel_count > 0 ) X { X printf " \n%6d articles were cancelled.\n", cancel_count X } X if ( precancel_count > 0 ) X { X printf " \n%6d articles were cancelled before receipt.\n", precancel_count X } X # X # Junkage report. Give total junkage, then break it down by X # reason. X # X printf "\n%6d articles were junked\n", junk_count X if ( junk_count > 0 ) X for ( group in junk_reason ) X printf " %6d %s\n", junk_reason[ group ], group X # X # Outgoing traffic report. Total ships, then break down by X # system. X # X printf( "\n%6d articles were shipped to other systems\n", ship_count ) X if ( ship_count > 0 ) X { X for ( to in ship_list ) X printf " %6d for %s\n", ship_list[ to ], to X printf " (Totals may differ due to same article shipped to multiple systems)\n" X } X # X # List how many articles we explictly rejected. Note we X # do not track the groups explicitly. X # X if ( x_count > 0 ) X { X printf "\n%6d articles were accepted but not posted/Xed by the active file\n", x_count X } X # X # Now we report on the questionable stuff X # X # X # This error comes first, as it indicates a severe problem X # with either you or your immediate neighbors X # X if ( no_msgid_count > 0 ) X { X printf "\n* * * Begin Serious Error! * * *\n" X printf "* * * There were attempts to insert articles which had *no* messages ids.\n" X printf "* * * This is a sign of significant errors in the posting or transfer\n" X printf "* * * software and should be checked out IMMEDIATELY!\n" X for ( source in no_msgid_source ) X printf "* * * %6d attempts were made from %s\n", no_msgid_source[ source ], source X printf "* * * End Serious Error! * * *\n" X } X # X # Other errors are less serious. X # X # The attempts to post to moderated groups. X # X printf "\n%6d articles were rejected as unapproved for moderated groups\n", unapproved_count X if ( unapproved_count > 0 ) X { X for ( target in unapproved_target ) X printf " %6d posted to %s\n", unapproved_target[ target ], target X for ( source in unapproved_source ) X printf " %6d came from %s\n", unapproved_source[ source ], source X } X # X # Now the transfer of unsubscribed stuff X # X printf "\n%6d articles were rejected as for unsubscribed groups\n", unsub_count X if ( unsub_count > 0 ) X { X for ( target in unsub_target ) X printf " %6d posted to %s\n", unsub_target[ target ], target X for ( source in unsub_source ) X printf " %6d came from %s\n", unsub_source[ source ], source X } X # X # List the articles rejected for having non-headers in the X # header section. Since the further spread of pedantic X # C news will eventually eliminate these, we do not report X # in the case where everything is OK. X # X if ( non_header_count > 0 ) X { X printf "\n%6d articles were rejected as having nonheaders in the header section\n", non_header_count X for ( source in non_header_source ) X printf " %6d came from %s\n", non_header_source[ source ], source X } X # X # List the number of articles rejected due to incorrect headers X # and where they came from. Since the further spread of pedantic X # C news will eventually eliminate these, we keep silent if X # everything is OK. X # X if ( bad_header_count > 0 ) X { X printf "\n%6d articles were rejected as having bad headers (no Date:, etc)\n", bad_header_count X for ( source in bad_header_source ) X printf " %6d came from %s\n", bad_header_source[ source ], source X } X # X # List those header fields and origins which were noted as X # being empty. Since the further spread of pedantic C news will X # eventually eliminate these, we keep silent if everything is OK. X # X if ( empty_header_count > 0 ) X { X printf "\n%6d articles were rejected as having empty headers\n", empty_header_count X for ( field in empty_header_field ) X printf " %6d entries had empty \"%s\" fields\n", empty_header_field[ field ], field X for ( source in empty_header_source ) X printf " %6d entries came from %s\n", empty_header_source[ source ], source X } X # X # Date rejections. Since the further spread of pedantic C news will X # eventually eliminate these, we keep silent if everything is OK. X # X if ( bad_date_count > 0 ) X { X printf "\n%6d articles were rejected as having unparseable dates:\n", bad_header_count X for ( source in bad_date_source ) X printf " %6d came from %s\n", bad_date_source[ source ], source X } X # X # These are article which are dated too far in the future. X # X if ( future_count > 0 ) X { X printf "\n%6d articles were rejected as dated too far in the future:\n", future_count X for ( source in future_source ) X printf " %6d came from %s\n", future_source[ source ], source X } X # X # Time rejects -- articles which are just too damned old. Probably X # signs of a news loop. X # X if ( oldies_count > 0 ) X { X printf "\n%6d articles were rejected as dated too far in the past:\n", oldies_count X for ( source in oldies_source ) X printf " %6d came from %s\n", oldies_source[ source ], source X } X # X # The fun stuff. This is where we report unrecognizable things. X # Data reported here is either meat for future mods to this script X # or indication of bugs in the news software. X # X if ( unknown_no_count > 0 ) X { X printf "\n%6d articles were rejected with an unrecognised comment\n", unknown_no_field X printf "about \"no such-and-such\" in the entry. Those rejections were:\n" X for ( no_field in unknown_no_field ) X printf " %s had the reason: %s\n", unknown_no_field[ no_field ], no_field X max_unknown = '$MAX_UNKNOWN' X printf "Here is up to %d of the entries:\n", max_unknown X for ( no_field in unknown_no_entry ) X if ( max_unknown-- > 0 ) X printf " %s\n", unknown_no_entry[ no_field ] X else X break X } X # X if ( nopath_count > 0 ) X { X printf "\n%6d articles were rejected with a missing Paths header.\n", nopath_count X max_nopath = '$MAX_NOPATH' X printf "Here are up to %d of the entries:\n", max_nopath X for ( nopath in nopath_entry ) X if ( max_nopath-- > 0 ) X printf " %s\n", nopath_entry X else X break X } X # X if ( mystery_dash_count > 0 ) X { X printf "\n%6d articles were rejected with an unrecognised \"-\" field.\n", mystery_dash_count X max_mystery_dash = '$MAX_MYSTERY_DASH' X printf "Here are up to %d of the entries:\n", max_mystery_dash X for ( mystery_dash in mystery_dash_entry ) X if ( max_mystery_dash-- > 0 ) X printf " %s\n", mystery_dash X else X break X } X if ( mystery_count > 0 ) X { X printf "\nThere were %d entries which were correctly formatted but with tag\n", mystery_count X printf "fields that were not recognized. Those tag fields and frequency were:\n" X tag_count = 0 X for ( mystery in mystery_tags ) X tag_count++ X for ( mystery in mystery_tags ) X { X if ( tag_count > 1 ) X printf " \"%s\" (%d),", mystery, mystery_tags[ mystery ] X else X printf " \"%s\" (%d)\n", mystery, mystery_tags[ mystery ] X tag_count-- X } X mystery_max='$MAX_MYSTERY' X printf "Here are up to %d of the entries:\n", mystery_max X for ( mystery in mystery_entry ) X if ( mystery_max-- > 0 ) X printf " %s\n", mystery X else X break X } X # X # Report on malformed lines in the log. Data reported here is X # almost certianly a bug in the news software. X # X defect_max = '$MAX_DEFECTS' X if ( defect_count > 0 ) X { X printf "\nThere were %d entries in the log with too few fields. Here is a sample:\n", defect_count X for ( defect in defective_entry ) X if ( defect_max-- > 0 ) X printf " \"%s\"\n", defect X else X break X } X } X # X # If nothing was wrong, print a nice reassuring message. X # X if ( ( mystery_count == 0 ) && ( mystery_dash_count == 0 ) && ( defect_count == 0 ) ) X printf "\nNo defects or unrecognized entries were found in the standard log.\n" X}' Xfi X/usr/ucb/Mail -s "News Log Report for $DATE" $NEWSMGR < $TEMPFILE END_OF_FILE if test 21554 -ne `wc -c <'cnews.logrep'`; then echo shar: \"'cnews.logrep'\" unpacked with wrong size! fi # end of 'cnews.logrep' fi echo shar: End of shell archive. exit 0 -- "If we don't provide support to our users someone is bound to confuse us with Microsoft." -- Charles "Chip" Yamasaki
) (05/27/91)
billd@fps.com (Bill Davidson) writes: > In article <57Lc310w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) > >Clearly, then, we have two alternatives: > >1. Report problems via some method other than each C News site sending mail; > >or > >2. Try to avoid problems. > >C News does neither. It therefore needs fixing. > > Actually, it does (1). It reposrts the problem in a log. This is not reporting the problem in any useful sense, since the report does not reach anyone who can do anything to fix it. I could set up our system to print out all mail bounces on an unattended printer sitting in an office in the Channel Islands, rather than mailing them back to the originator. This would not generally be construed as "reporting the problem", however. mathew
) (05/27/91)
chip@tct.com (Chip Salzenberg) writes: > You don't realize the importance of the Date header. Older C News > versions did as you suggest. As a result, they were vulnerable to the > loads of old news which get dumped on the net periodically due to > admin inexperience and mistakes. C News insists on a parseable Date > so it can guarantee that the news it forwards isn't old junk. Yes. The problem is that this "guarantee", along with the other "guarantees" implemented by C News, has the side effect that ordinary people can have their postings silently dropped because they made trivial typos. Which is more valuable -- the "guarantee" that no old news will reappear, or the "guarantee" that people's postings will get out to the net? I'd say the latter. mathew
) (05/27/91)
rickert@mp.cs.niu.edu (Neil Rickert) writes: > In article <1991May26.082113@mccall.com> tp@mccall.com (Terry Poot) writes: > >the RFC's before doing so. Are all your users required to read RFC822 and > >RFC1036 before using news? Mine aren't, and I don't think they should be. > > I don't think users should be required to read RFC1036 either. But I do > think news software developers should, and should make sure that articles > entered into the news system with their software are compliant. Right. Now supposing they don't. Or supposing that, being only human, they make a mistake (GASP!) What's the result? Innocent bystanders get punished by having their articles silently deleted. If you think those innocent bystanders aren't going to hold C News at least partially to blame, then I think you're naive. And if you think they aren't going to turn up in this newsgroup every few weeks or so and start a major flame war over this issue, then I think you're naive. Do you really want that? mathew
Paul Ribeiro <ACAD8012@Ryerson.Ca> (05/27/91)
In article <1cXL327w164w@mantis.co.uk>, mathew@mantis.co.uk (CNEWS MUST DIE!) says: >Right. Now supposing they don't. Or supposing that, being only human, they >make a mistake (GASP!) then they fix it..seems to me you are the one thats broken, not C News :-)
rickert@mp.cs.niu.edu (Neil Rickert) (05/28/91)
In article <1cXL327w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >If you think those innocent bystanders aren't going to hold C News at least >partially to blame, then I think you're naive. And if you think they aren't Some users probably will blame C news. This is unfortunate. For, as long as they insist on ranting and raving at a program which is not the cause of their problems, they will only needlessly postpone the search for a real solution. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
rickert@mp.cs.niu.edu (Neil Rickert) (05/28/91)
In article <qZwL326w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >Yes. The problem is that this "guarantee", along with the other "guarantees" >implemented by C News, has the side effect that ordinary people can have >their postings silently dropped because they made trivial typos. > >Which is more valuable -- the "guarantee" that no old news will reappear, or >the "guarantee" that people's postings will get out to the net? I'd say the >latter. Speaking as the admin of a site very short on spare disk space, the guarantee that old news won't reappear is more important. But then, as a site admin, I accept the responsibility of ensuring that postings by local users are not silently dropped elsewhere. The high quality standards of the newest Cnews helps me live up to that responsibility. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
sef@kithrup.COM (Sean Eric Fagan) (05/28/91)
In article <qZwL326w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >Yes. The problem is that this "guarantee", along with the other "guarantees" >implemented by C News, has the side effect that ordinary people can have >their postings silently dropped because they made trivial typos. If ordinary people are giving the Date: line to a posting, something is broken with their software, anyway. Since a program is supposed to spit out the date, and I think H&G even supply a routine to do so if you don't trust the one that comes with your system, anyone not following the RFC deserves, as several people (nearly everyone but you, it seems) have pointed out, to have their postings dropped on the floor. >Which is more valuable -- the "guarantee" that no old news will reappear, or >the "guarantee" that people's postings will get out to the net? I'd say the >latter. I'd say the former. -- Sean Eric Fagan | "I made the universe, but please don't blame me for it; sef@kithrup.COM | I had a bellyache at the time." -----------------+ -- The Turtle (Stephen King, _It_) Any opinions expressed are my own, and generally unpopular with others.
randy@m2xenix.psg.com (Randy Bush) (05/28/91)
mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >> Actually, it does (1). It reposrts the problem in a log. > This is not reporting the problem in any useful sense, since the report does > not reach anyone who can do anything to fix it. You mean that you run software that produces an error log and don't run that log into mail for the SA nightly? If so, then maybe this discussion should move to news.newusers.questions or alt.bbs.waffle. -- randy@psg.com ..!uunet!m2xenix!randy
tneff@bfmny0.BFM.COM (Tom Neff) (05/28/91)
In article <17898@celit.fps.com> billd@fps.com (Bill Davidson) writes: >In article <1991May20.204313.25340@decuac.dec.com> mjr@hussar.dco.dec.com (Marcus J. Ranum) writes: >>I personally don't care if my server spends 5% more time a day processing >>news, as long as I don't have to spend any more time dealing with it. > >Some of us do. Some of us are lucky to get a computer that can handle >the load even under ideal software conditions. Given that load tends to expand to consume available resources, it is not surprising that a lot of people are already maxed out -- but it means that it's pointless to worry about whether everyone's going to be able to afford a 5% cycle hike for some useful new feature. There will always be people who can't, or have convinced themselves they can't, whether rationally or irrationally. In my experience, 3/4 of the 'maxed out' sites in the world are doing at least one thing so STUPIDLY that they could free up giant chunks of capacity with a stroke, if only they knew what they were doing. I am sure Bill Davidson is not one of these; yet even in the case where a smart admin is managing a system at capacity, there is a tradeoff you can outline. Is building a more error free network worth sacrificing X percent of your throughput every day? If not, exactly WHY not? If some rampaging news explosion that *could* have been prevented by additional "insurance" software -- at a 1% cost per day -- succeeds in knocking your cluster out for a weekday, was it worth it? You can buy your way out of a capacity bind, but not (in general) a reliability bind.
randy@m2xenix.psg.com (Randy Bush) (05/28/91)
mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >> I don't think users should be required to read RFC1036 either. But I do >> think news software developers should, and should make sure that articles >> entered into the news system with their software are compliant. > Right. Now supposing they don't. Or supposing that, being only human, they > make a mistake (GASP!) > What's the result? Innocent bystanders get punished by having their articles > silently deleted. The alternative is for all news software to attempt to *correct* all possible errors. This problem is not solvable, and is not very interesting either. In the words of Erik Naggum on the subject of net software RFC compliance, "It must be our great task, gentlemen, to keep the monkeys away from the typewriters." C News is a positive step in that direction. Thanks Geoff and Henry. > If you think those innocent bystanders aren't going to hold C News at least > partially to blame, then I think you're naive. And if you think they aren't > going to turn up in this newsgroup every few weeks or so and start a major > flame war over this issue, then I think you're naive. Expected. The typewriters are very attractive. But for us to hear your screaming, you will have to use non-broken software. Goal accomplished. > Do you really want that? We all have ^K, and use it when necessary. Occasionally one reads the stuff for amusement. Pretty booooring, if you ask me, and you did. But a coupla decades back we spent three years in California and learned about 'different strokes'. So, if screaming gets you off, enjoy. -- randy@psg.com ..!uunet!m2xenix!randy
nreadwin@micrognosis.co.uk (Neil Readwin) (05/28/91)
In article <qZwL326w164w@mantis.co.uk>, CNEWS MUST DIE! <mathew@mantis.co.uk> writes: |> Which is more valuable -- the "guarantee" that no old news will reappear, or |> the "guarantee" that people's postings will get out to the net? Let me put it this way. If somebody dumps old news onto the net then it wastes a lot of my time - either rereading things or (in extreme cases) repairing the damage. If (say) your postings get dumped then I *save* time since I don't get to read them. In fact if somebody wrote a news transport that dumped every article with a prime number in the message id but guaranteed not to flood the net with old postings then I would still pick the former. Phone: +44 71 528 8282 E-mail: nreadwin@micrognosis.co.uk Everything is a cause for sorrow that my mind or body has made
skrenta@blekko.commodore.com (Rich Skrenta) (05/28/91)
mathew@mantis.co.uk (CNEWS MUST DIE!) writes: > If you think those innocent bystanders aren't going to hold C News at least > partially to blame, then I think you're naive. The ignorant can blame whoever or whatever they like without concern for whether or not they have the slightest clue what they're talking about. > And if you think they aren't > going to turn up in this newsgroup every few weeks or so and start a major > flame war over this issue, then I think you're naive. So you're saying that until Cnews gets the modifications YOU want we'll be sorry because you and other whiners will fill news.admin with crap until you get what you want? Get real. > Right. Now supposing they don't. Or supposing that, being only human, they > make a mistake (GASP!) Psst. Gotta neat trick for you. You can stick a little signature in the Message-ID header. Put a message before the '@', and everyone will see it. Always use the same little message. Rich -- skrenta@blekko.commodore.com
scs@iti.org (Steve Simmons) (05/28/91)
mathew@mantis.co.uk (CNEWS MUST DIE!) raves: >rickert@mp.cs.niu.edu (Neil Rickert) writes: >> I don't think users should be required to read RFC1036 either. But I do >> think news software developers should, and should make sure that articles >> entered into the news system with their software are compliant. >Right. Now supposing they don't. Or supposing that, being only human, they >make a mistake (GASP!) Only reasonable. Why put the date in the article for users to muck with? The posting software should either get it right or fix it, not dump bad format news into the transport mechanism. >What's the result? Innocent bystanders get punished by having their articles >silently deleted. The posting software should catch such errors and inform the user while there is still a chance to fix it. The posting software should not dump bad news into the transport mechanism. >If you think those innocent bystanders aren't going to hold C News at least >partially to blame, then I think you're naive. And if you think they aren't >going to turn up in this newsgroup every few weeks or so and start a major >flame war over this issue, then I think you're naive. The user should blame the software that allowed me to post the bad date in the first place. The posting software should not dump bad news into the transport mechanism. As for flame wars, some *ssh*l* will turn up in this group and start one without the slightest regard for facts. -- "If we don't provide support to our users someone is bound to confuse us with Microsoft." -- Charles "Chip" Yamasaki
scs@iti.org (Steve Simmons) (05/28/91)
scs@iti.org (Steve Simmons) writes: >The following script generates a detailed report on news traffic and >problems including header reports. It's by no means done, but it >seems clear there's a need...the folks who've been beta testing it >seem to like it. Patches will be appreciated, but be forewarned it's >going to get a major rewrite Real Soon Now. Several folks have reported that awk chokes on the large internal script. That's one of the reason's it's going to get rewritten. Breaking the awk out to a separate file will probably do the fix. -- "If we don't provide support to our users someone is bound to confuse us with Microsoft." -- Charles "Chip" Yamasaki
) (05/28/91)
sef@kithrup.COM (Sean Eric Fagan) writes: > In article <qZwL326w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) > >Yes. The problem is that this "guarantee", along with the other "guarantees" > >implemented by C News, has the side effect that ordinary people can have > >their postings silently dropped because they made trivial typos. > > If ordinary people are giving the Date: line to a posting, something is > broken with their software, anyway. C News will also, I gather, complain about Keywords:foo,bar > anyone not following the RFC deserves, > as several people (nearly everyone but you, it seems) have pointed out, to > have their postings dropped on the floor. Right. But the users who happen to be using the software written by the person who didn't follow the RFC are innocent bystanders; and innocent bystanders should not be punished. > >Which is more valuable -- the "guarantee" that no old news will reappear, or > >the "guarantee" that people's postings will get out to the net? I'd say the > >latter. > > I'd say the former. Then you presumably have a rather strange idea of what the purpose of Usenet is. mathew
) (05/28/91)
randy@m2xenix.psg.com (Randy Bush) writes: > mathew@mantis.co.uk (CNEWS MUST DIE!) writes: > > What's the result? Innocent bystanders get punished by having their articl > > silently deleted. > > The alternative is for all news software to attempt to *correct* all possible > errors. OR REPORT THEM. mathew
karish@mindcraft.com (Chuck Karish) (05/29/91)
In article <qZwL326w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >Which is more valuable -- the "guarantee" that no old news will reappear, or >the "guarantee" that people's postings will get out to the net? I'd say the >latter. The "guarantee" that no old news will reappear is one that news hosts demand of each other. An arbitrary doubling of news volume, as happens when there's a flood of old articles, is intolerable. If the site at fault were to fail to correct such a problem, it would be cut off by its neighbors. The "guarantee" that postings will be propagated is one that's made by news software writers and sytem managers who maintain their systems in conformance with RFC1036, which defers to RFC822 as the authority for date formats. If you don't follow the standards, you're on your own, and you've jeopardized your guarantee to your users. -- Chuck Karish karish@mindcraft.com Mindcraft, Inc. (415) 323-9000
rickert@mp.cs.niu.edu (Neil Rickert) (05/29/91)
In article <wu2N39w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >Right. But the users who happen to be using the software written by the >person who didn't follow the RFC are innocent bystanders; and innocent >bystanders should not be punished. I totally agree. That is why we are tired at your continued attempts to harrass the totally innocent designers of Cnews, and the totally innocent sysadmins who have made an honest and considered judgement to use Cnews. Why don't you go attack the software which does the posting instead. There you have the real cause of your problem. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
scott@skypod.guild.org (Scott Campbell) (05/29/91)
In article <1991May21.155936.9037@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes: >There may not be one "first C News site" to see the article; there may be >dozens or even hundreds of them, if the article first passed through a >B News site with a large news fanout. ...like uunet? -- Scott J.M. Campbell scott@skypod.guild.org Skypod Communications Inc. ..!gatech!dscatl!daysinns!skypod!scott 1001 Bay Street, Suite 1210 ..!uunet!utai!lsuc!becker!skypod!scott Toronto, Ont. (416) 924-4059 ..!epas.utoronto.ca!nyama!skypod!scott
jfw@ksr.com (John F. Woods) (05/29/91)
mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >sef@kithrup.COM (Sean Eric Fagan) writes: >>>Which is more valuable -- the "guarantee" that no old news will reappear, or >>>the "guarantee" that people's postings will get out to the net? I'd say the >>>latter. >> >>I'd say the former. >Then you presumably have a rather strange idea of what the purpose of Usenet >is. And you have (a) a rather quaint notion of just how "little" old news there is, relative to the carrying capacity of poor, overworked little modems, and (b) a curious notion of the overwhelming importance of 99 44/100 % of what gets typed to USENET...
kherron@ms.uky.edu (Kenneth Herron) (05/29/91)
mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >C News will also, I gather, complain about Keywords:foo,bar >> anyone not following the RFC deserves, >> as several people (nearly everyone but you, it seems) have pointed out, to >> have their postings dropped on the floor. >Right. But the users who happen to be using the software written by the >person who didn't follow the RFC are innocent bystanders; and innocent >bystanders should not be punished. The headers accessible to the user need not be the same headers sent out to the world. The posting software is in a fine position to either fix the headers or bounce the article with a *meaningful* error message to the *correct* person. Why then, do you want it to pass the buck to software on another site? If the originating site makes sure the headers are right, then it has nothing to worry about. If it releases an article with an invalid headers, then it dropped the ball, not the user. Why should the transport agent fix up the article if the posting software couldn't be bothered to? Could it be because *you* wrote the posting software in question? -- Kenneth Herron kherron@ms.uky.edu University of Kentucky +1 606 257 2975 Department of Mathematics "So this won't be a total loss, can you make it so guys get to throw their mothers-in-law in?" "Sure, why not?"
aipdc@castle.ed.ac.uk (Paul D. Crowley) (05/29/91)
In article <3752@ksr.com> jfw@ksr.com (John F. Woods) writes: |mathew@mantis.co.uk (CNEWS MUST DIE!) writes: |>sef@kithrup.COM (Sean Eric Fagan) writes: |>>>Which is more valuable -- the "guarantee" that no old news will reappear, or |>>>the "guarantee" that people's postings will get out to the net? I'd say the |>>>latter. |>> |>>I'd say the former. |>Then you presumably have a rather strange idea of what the purpose of Usenet |>is. | |And you have (a) a rather quaint notion of just how "little" old news there |is, relative to the carrying capacity of poor, overworked little modems, and |(b) a curious notion of the overwhelming importance of 99 44/100 % of what |gets typed to USENET... I'm convinced you've all gone mad. I use MicroEMACS as an editor. If I try to exit without saving an altered file, it asks me if I really want to do that, and sometimes I don't. When I send mail, that mail is bounced back to me if I get the address wrong, and I get a chance to edit it and try again. When I type rm *, I get asked "Are you sure?" If I send news with an incorrect date header, it is silently dumped. No, I tell a lie, it is not silent at all. An error is written in the log file of the dumping site. "Didn't you see the notice?" "Yes, I did. It was at the bottom of a locked filing cabinet in a disused lavoratory with a sign on the door saying 'Beware of the Leopard!'. Ever thought of going into advertising?" --- Douglas Adams, "The Hitch-Hikers Guide to the Galaxy". I don't think there's anything wrong with the attitude "Unfortunately, it's impractical: a thousand sites could try to send errors all at once, and that wouldn't be good." If it's unavoidable behaviour, no problem. The trouble is, no-one wants to look for ways around it, because they feel it's _desirable_ behaviour, to "punish" the user for making a mistake. I can't see why news is an exception to the general rule that software is supposed to be _nice_ to the user, and not deliberately spiteful. Various people don't seem interested in discussing alternatives such as the "control" hack because they get a kick out of the idea of spiteful software -- so long as it doesn't happen to them. At the heart of it is this: |(b) a curious notion of the overwhelming importance of 99 44/100 % of what |gets typed to USENET... (b) isn't an argument for the dumping of random arguments. It's an argument for the abolition of Usenet. If you don't buy it, it's probably because you have some idea of how useful news can be. Look at comp.windows.x, comp.lang.c++, sci.virtual-worlds, and you see discussion which is actually forwarding the craft, and not just for the amusement of the participants. Junk articles every so often out of spite and that usefulness is impaired. Sure, the world would be a better place if software never had any bugs in it, if human users never made mistakes, if everyone understood and adhered to the same standards. Given that they don't, isn't it part of writing half-decent software taking into account the possibility that not everything and everyone you're taking data from is entirely without flaw? ____ \/ o\ Paul Crowley aipdc@castle.ed.ac.uk \ / /\__/ Part straight. Part gay. All queer. \/ "I say we kill him and eat his brain." "That's not the solution to _every_ problem, you know!" -- Rudy Rucker
rdc30@nmrdc1.nmrdc.nnmc.navy.mil (LCDR Michael E. Dobson) (05/29/91)
In article <1991May19.013616.10567@ms.uky.edu> kherron@ms.uky.edu (Kenneth Herron) writes: >chip@tct.com (Chip Salzenberg) writes: > >>>> News software which generates articles that do not conform to the RFC is >>>> broken. Articles generated by such buggy software deserve to be discarded. > >One problem with this is that the C-news report only lists the site(s) >which are feeding you bad articles, not where the articles actually >came from or the exact problem with the article. The message IDs are logged >along with the complaint , but the messages themselves are thrown away >(of course) so the concerned administrator has to do a bit of detective work. > grep 'site -' log.o > munged.list gave me the actual message id's of the site generating the article (once I filtered out the duplicate entries). I could then send e-mail to the offending site. I did this for each feed when I was getting lots of bad articles, now that it's slowed to a trickle, it't even easier. >I'm working on a script to notify me of all articles bit-bucketed for >bad header reasons. I could arrange for it to be posted, e.g. to >news.lists. Is there any call for such a thing? You could also post to news.software.b -- Mike Dobson, Sys Admin for | Internet: rdc30@nmrdc1.nmrdc.nnmc.navy.mil nmrdc1.nmrdc.nnmc.navy.mil | UUCP: ...uunet!mimsy!nmrdc1!rdc30 AT&T 3B2/600G Sys V R 3.2.2 | BITNET: dobson@usuhsb or nrd0mxd@vmnmdsc WIN/TCP for 3B2 | MCI-Mail: 377-2719 or 0003772719@mcimail.com
darcy@druid.uucp (D'Arcy J.M. Cain) (05/30/91)
In article <10623@castle.ed.ac.uk> Paul D. Crowley writes: >The trouble is, no-one wants to look for ways around it, because they >feel it's _desirable_ behaviour, to "punish" the user for making a >mistake. I can't see why news is an exception to the general rule that >software is supposed to be _nice_ to the user, and not deliberately >spiteful. Various people don't seem interested in discussing >alternatives such as the "control" hack because they get a kick out of >the idea of spiteful software -- so long as it doesn't happen to them. It doesn't happen to them because they are running software that checks the message before sending it out to the cold hard world of Usenet connectivity. Of course the user should be coddled. No one that I have seen supports the notion of unfriendly user interfaces. The argument is that the software that interfaces with the user is the place where this stuff should be checked. If you are writing posting software you should check everything the user enters to make sure that it conforms to the relevant specifications. No one on the "CNEWS IS RIGHT" side suggests otherwise. It's the "CNEWS MUST BE WHINED AT AND DIE" group that thinks that the posting software can slough that stuff onto software running 6 or 7 hops away. That's like arguing that print spoolers should correct spelling because users make spelling mistakes. So, the $64K question: Who or what should take responsibility for the correctness of news headers? A) The novice poster that probably hasn't even heard of an RFC. B) The software that injects the message into the net. C) The software that transports the message around the world. CNews suggests that the right answer is B. It's detractors claim that C is right but they seem to argue that the only other alternative is A. So, does anyone want to admit that they believe that the right answer is A? If not can we at least agree that the choice is either B or C and admit that no one is suggesting that the user should be making the final determination on headers before the message enters the net? Now, can we at least all decide that the right choice is either B or C or is there anyone out there arguing that the user (apart from his/her choice of posting software) is not the right place to have headers -- D'Arcy J.M. Cain (darcy@druid) | D'Arcy Cain Consulting | There's no government Toronto, Ontario, Canada | like no government! +1 416 424 2871 |
henry@zoo.toronto.edu (Henry Spencer) (05/30/91)
In article <1991May28.213321.17039@skypod.guild.org> scott@skypod.guild.org (Scott Campbell) writes: >>There may not be one "first C News site" to see the article; there may be >>dozens or even hundreds of them, if the article first passed through a >>B News site with a large news fanout. > >...like uunet? That's the example I had in mind, although it's now obsolete since uunet is running C News now. -- "We're thinking about upgrading from | Henry Spencer @ U of Toronto Zoology SunOS 4.1.1 to SunOS 3.5." | henry@zoo.toronto.edu utzoo!henry
david@golem.uucp (David J. Fiander) (05/30/91)
In article <10623@castle.ed.ac.uk> aipdc@castle.ed.ac.uk (Paul D. Crowley) writes: >I'm convinced you've all gone mad. > >When I send mail, that mail is bounced back to me if I get the address >wrong, and I get a chance to edit it and try again. > >If I send news with an incorrect date header, it is silently dumped. >No, I tell a lie, it is not silent at all. An error is written in the >log file of the dumping site. The first several examples you gave (of which I have only kept two) are all very good examples of what should happen when a user does something he may not want to. The most telling point is the mail example. Since mail is point-to-point, you will only get one bounce (unless your name is Eric Fair), but that is not feasible for news. As well, address problems cannot necessarily be determined on your machine, but you mail system probably does make sure that the headers that it transmits to other sites are syntactically correct. The problem is that the news software on _your_ system should notice that something is the matter with the article headers that you have provided and either correct or bounce the message. CNEWS MUST DIE has been complaining that his NUA (generalising from the mail world causes us to have a News User Agent), allows him to create bogus messages, and that he doesn't find out when somebody else drops it. As several other people have pointed out, the "correct" solution to Mr. DIE's problem is to fix his software so that it will not accept malformed messages. If you post a malformed article on a node running Cnews, then Cnews will fix the headers so that they are standard-conforming (Henry, will it bounce a message back to the originator if it's totally gone?). All the flames have been about the fact that Cnews won't be as polite about messages that are arriving from remote sites. -- David J. Fiander <david@golem.uucp> L'homme est bien insens\'e. Il ne saurait forge un ciron, et forge des Dieux \`a douzaines. - Montaigne (1533 -- 1592)
chip@tct.com (Chip Salzenberg) (05/30/91)
According to aipdc@castle.ed.ac.uk (Paul D. Crowley): >I don't think there's anything wrong with the attitude "Unfortunately, >it's impractical: a thousand sites could try to send errors [...]" > >The trouble is, no-one wants to look for ways around it, because they >feel it's _desirable_ behaviour, to "punish" the user for making a >mistake. I consider it desirable to put teeth in the standards for *software* that transports news. Each person who ventures to write Usenet software enters a field in which small mistakes can cause huge problems. To require the relatively few people who write the software to adhere *strictly* to the relevant RFCs is, in my opinion, not too much to ask. To users, I say: Choose carefully the software and the system you use to access Usenet. If you choose badly, you suffer the consequences. >Sure, the world would be a better place if software never had any bugs >in it, if human users never made mistakes, if everyone understood and >adhered to the same standards. Given that they don't, isn't it part of >writing half-decent software taking into account the possibility that >not everything and everyone you're taking data from is entirely without >flaw? C News matches that description. It antipates and recovers from errors quite well, thank you. The only point of disagreement here is what kind of error recovery is appropriate. If robustness were the only issue, Geoff and Henry would be universally acclaimed. -- Chip Salzenberg at Teltronics/TCT <chip@tct.com>, <uunet!pdn!tct!chip> perl -e 'sub do { print "extinct!\n"; } do do()'
dlr@daver.bungi.com (Dave Rand) (05/30/91)
In article <1991May29.221015.1865@druid.uucp> darcy@druid.uucp (D'Arcy J.M. Cain) writes: >So, the $64K question: Who or what should take responsibility for >the correctness of news headers? > A) The novice poster that probably hasn't even heard of an RFC. > B) The software that injects the message into the net. > C) The software that transports the message around the world. I run an email-to-news gateway of the pc532 mailing list. Up until today I was silently ignoring this thread, thinking that I was ok. Turns out I wasn't. No big deal - I fixed my posting software by changing to the r$ provided flavour. If you have a problem, fix it at the source. Don't depend on others to do the work for you. Cnews is _NOT_ broken, it is enforcing good and proper restrictions. I like it. -- Dave Rand {pyramid|mips|bct|vsi1}!daver!dlr Internet: dlr@daver.bungi.com
aipdc@castle.ed.ac.uk (Paul Crowley) (05/30/91)
In article <1991May29.221015.1865@druid.uucp> darcy@druid.uucp (D'Arcy J.M. Cain) writes: >That's like arguing that print spoolers should correct spelling because >users make spelling mistakes. No, it's like arguing that if a print spooler is going to junk files with spelling mistakes it should at least tell the user about them. >So, the $64K question: Who or what should take responsibility for >the correctness of news headers? I'm sorry I didn't make myself clear in my original post. This isn't the $64k question at all, and the answer to your strawman question is obviously B. The $64k question is: Who or what should behave sensibly when it sees an incorrect header? > A) The novice poster that probably hasn't even heard of an RFC. > B) The software that injects the message into the net. > C) The software that transports the message around the world. And the correct answer is: A. B. C. ____ \/ o\ Paul Crowley aipdc@castle.ed.ac.uk \ / /\__/ Part straight. Part gay. All queer. \/ "I say we kill him and eat his brain." "That's not the solution to _every_ problem, you know!" -- Rudy Rucker
) (05/30/91)
nreadwin@micrognosis.co.uk (Neil Readwin) writes: > Let me put it this way. If somebody dumps old news onto the net then it > wastes a lot of my time - either rereading things or (in extreme cases) > repairing the damage. If (say) your postings get dumped then I *save* time > since I don't get to read them. Ah. So the purpose of Usenet is to save time? In that case, why not unplug your modem? Think of the time you'll save then! > In fact if somebody wrote a news transport > that dumped every article with a prime number in the message id but guarantee > not to flood the net with old postings then I would still pick the former. Let me ask you a question: Why do you bother to get Usenet if you have so little interest in its content? mathew
) (05/30/91)
rickert@mp.cs.niu.edu (Neil Rickert) writes: > In article <wu2N39w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) w > >Right. But the users who happen to be using the software written by the > >person who didn't follow the RFC are innocent bystanders; and innocent > >bystanders should not be punished. > > I totally agree. That is why we are tired at your continued attempts to > harrass the totally innocent designers of Cnews, and the totally innocent > sysadmins who have made an honest and considered judgement to use Cnews. But C News is the program which is meting out the punishment! Or is it "only following orders", as the famous saying goes? mathew
henry@zoo.toronto.edu (Henry Spencer) (05/30/91)
In article <1991May30.015111.375@golem.uucp> david@golem.uucp (David J. Fiander) writes: >If you post a malformed article on a node running Cnews, then Cnews >will fix the headers so that they are standard-conforming (Henry, will >it bounce a message back to the originator if it's totally gone?). Yup. -- "We're thinking about upgrading from | Henry Spencer @ U of Toronto Zoology SunOS 4.1.1 to SunOS 3.5." | henry@zoo.toronto.edu utzoo!henry
) (05/30/91)
darcy@druid.uucp (D'Arcy J.M. Cain) writes: > In article <10623@castle.ed.ac.uk> Paul D. Crowley writes: > > Various people don't seem interested in discussing > >alternatives such as the "control" hack because they get a kick out of > >the idea of spiteful software -- so long as it doesn't happen to them. > > It doesn't happen to them because they are running software that checks > the message before sending it out to the cold hard world of Usenet > connectivity. Of course the user should be coddled. No one that > I have seen supports the notion of unfriendly user interfaces. The > argument is that the software that interfaces with the user is the > place where this stuff should be checked. If you are writing posting > software you should check everything the user enters to make sure > that it conforms to the relevant specifications. No one on the "CNEWS > IS RIGHT" side suggests otherwise. It's the "CNEWS MUST BE WHINED AT > AND DIE" group that thinks that the posting software can slough that > stuff onto software running 6 or 7 hops away. Where the hell do you get that idea? I am *not* saying that posting software should be allowed to produce bad articles. I agree entirely that broken software should be fixed. My argument is that there will always be bad articles which get past the posting software. There is such a diversity of posting software that this is inevitable. Also, any software which allows the user to edit headers is asking for trouble in this respect. I fixed the broken software I was using as soon as I found out about the error. I would expect anyone else to do the same. THE POINT is that I didn't find out about the error for weeks, because C News made no attempt to tell me that I had failed its unnecessarily stringent requirements. There are also the additional points that: - if correcting an error is possible, it should be corrected; - if *you* can safely ignore an error, you should do so. > That's like arguing > that print spoolers should correct spelling because users make > spelling mistakes. No, it's like arguing that Postscript printers should report an error when someone sends them an ASCII file, rather than just silently losing the whole of the input and leading the user to wonder what exactly is going on. We have a printer here at work which silently junks any page containing bitmap data sent in the wrong format. I think that said printer is profoundly badly designed. > So, the $64K question: Who or what should take responsibility for > the correctness of news headers? > A) The novice poster that probably hasn't even heard of an RFC. > B) The software that injects the message into the net. > C) The software that transports the message around the world. > > CNews suggests that the right answer is B. It's detractors claim that > C is right but they seem to argue that the only other alternative is A. No, C News suggests that the ONLY answer is B; that if for some reason a bad message has got past the posting software, then that message is fair game for deletion. I am claiming that the answer is B *and* C. > So, does anyone want to admit that they believe that the right answer is > A? If not can we at least agree that the choice is either B or C Nope, I don't think you should be using exclusive-or there. mathew
darcy@druid.uucp (D'Arcy J.M. Cain) (05/31/91)
In article <10674@castle.ed.ac.uk> Paul Crowley writes: >In article <1991May29.221015.1865@druid.uucp> darcy@druid.uucp (D'Arcy J.M. Cain) writes: >>So, the $64K question: Who or what should take responsibility for >I'm sorry I didn't make myself clear in my original post. This isn't >the $64k question at all, and the answer to your strawman question is >obviously B. The $64k question is: Who or what should behave sensibly >when it sees an incorrect header? >> A) The novice poster that probably hasn't even heard of an RFC. >> B) The software that injects the message into the net. >> C) The software that transports the message around the world. >And the correct answer is: >A. B. C. I'll buy that. However I suspect we disagree on what the sensible thing for transport software to do is. You seem to think it should start a flood of reports back to the originator. I think that CNews does the sensible thing and drops them. So let me prop up my straw man a little more. (Ahh, if I only had a brain ...) 1 - Anyone who posts to the net should make a reasonable effort to get the headers right. The newsgroup line should reflect the subject matter as should the subject. If a thread is getting old perhaps trim the references line. Make the follow-up line and the distribution reasonable. The poster should try to get the format of the lines right but they shouldn't be the last defense on this point. 2 - The posting software should not try to decide whether the subject reflects the content or whether the cross-posting is relevant. It should however make sure that every header line follows the RFCs. If not it should at least notify the user immediately that there is a problem. It might also try to correct the situation with or without the user's help depending on the nature of the problem. 3 - The transport software should move news through a variety of networks. To do this efficiently it should act on a well defined set of rules. Messages that fail to conform to these rules should not be passed. Now I really doubt that many people will have any substantial argument with #1. So who disagrees with #2. (Don't post if you agree. Of course you do. You'll just flood the newsgroup.) Now if you disagree with #2 and agree with #1 then you are inexorably drawn to the conclusion that #3 is wrong and that the news transport software should be fixing the user's mistakes. I doubt you can be convinced otherwise. So if you agree with #1 and #2 and disagree with #3, on what basis do you do so? Remember that the user is protected by the posting software so buttinsky transport software is only protecting bad posting software. I happen to believe that if you are writing software that injects news into the news system than the onus is on you to follow the standards and not to expect Henry and Geoff to conform to yours. Go ahead. I still have an I, II and III waiting in the wings. :-) -- D'Arcy J.M. Cain (darcy@druid) | D'Arcy Cain Consulting | There's no government Toronto, Ontario, Canada | like no government! +1 416 424 2871 |
darcy@druid.uucp (D'Arcy J.M. Cain) (05/31/91)
In article <T6mR328w164w@mantis.co.uk> CNEWS MUST DIE! writes: >But C News is the program which is meting out the punishment! >Or is it "only following orders", as the famous saying goes? What orders? You seem to be the one with the orders and Cnews certainly isn't following them. -- D'Arcy J.M. Cain (darcy@druid) | D'Arcy Cain Consulting | There's no government Toronto, Ontario, Canada | like no government! +1 416 424 2871 |
louie@sayshell.umd.edu (Louis A. Mamakos) (05/31/91)
In article <7moR331w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >I fixed the broken software I was using as soon as I found out about the >error. I would expect anyone else to do the same. THE POINT is that I didn't >find out about the error for weeks, because C News made no attempt to tell me >that I had failed its unnecessarily stringent requirements. Just two pointa after following this wonderful discussion: * What the #$^&* do you want for free? Just what does anyone pay for cnews that justifies demands upon the authors to change fundmaental design decisions on their part? cnews works better than some software that I've actually paid money for. * If you feel so strongly about the problems that cnews has, why not fix it yourself, and offer to distribute your version to sites on the net that are run by folks that agree with your point of view. You certainly seem to have the time to do this based upon the volume of your articles that seem to say pretty much the same thing over and over again. I know how difficult it is to package up software in a form suitable for distribution. We should be thankful that Henry and Geoff have gone to this much trouble to share their effort with the rest of us. I suppose that if you don't like the way this USENET is run [by consensus, sort of], then go off and start your own. I for one am tired of hearing pot shots at positions and opinions that have been clearly stated on many occasions.
nreadwin@micrognosis.co.uk (Neil Readwin) (05/31/91)
In article <J2mR327w164w@mantis.co.uk>, CNEWS MUST DIE! <mathew@mantis.co.uk> writes: |> Let me ask you a question: Why do you bother to get Usenet if you have so |> little interest in its content? I am very interested in the contents of most of the Usenet groups I read. That is why I read them. Sorting out things when the disk fills up is not interesting, I do it because I have to. Rereading articles I read two months ago is not interesting either. Losing some proportion of the stuff posted to a group is not a problem, the group remains equally interesting. So ... I still say that I would rather lose news than have old new recirculated. Phone: +44 71 528 8282 E-mail: nreadwin@micrognosis.co.uk Everything is a cause for sorrow that my mind or body has made
mrm@sceard.Sceard.COM (M.R.Murphy) (05/31/91)
In article <T6mR328w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: > >But C News is the program which is meting out the punishment! > >Or is it "only following orders", as the famous saying goes? mathew, this is the way you do it. 0) You decide you want to use USENET. 1) You install News software. 2) You post an article with a distribution limited either by the Distribution: header or the Newsgroups: header so that the article will be seen only at your local site. You do this to avoid looking the fool. You inspect the article that was posted with the tools that you have for looking at articles. This includes a text editor that doesn't hide headers. You look at logs. You look for errors. 3) You repeat 2) until you are satisfied that your postings are not pond scum. 4) You post an article with a distribution limited either by the Distribution: header or the Newsgroups: header so that the article will be seen only at your local site and a single site that you feed. This is usually done by posting to a newsgroup like "to.othersite". You inspect the article that was posted with the tools that you have for looking at articles. You look at logs. You look for errors. You ask that the administrator at the site that you feed do the same. 5) You repeat 4) until you are satisfied that your postings are not pond scum. 6) You repeat 4) and 5) for each of the sites that you feed. 7) You post to a regional test group or to misc.test with limited distribution You phrase your posting so that it asks for help rather than demands help. 8) You wait for some responses from autoresponse daemons and friendly administrators. 9) If you don't get responses, you ask the sites you feed and that feed you for help. Nicely. 10) If you change your software or add a feed, you repeat 2)-9), as required. 11) If you want to know how far you get, post to misc.test with an unlimited distribution. You are allowed to do this with reasonable frequency to see if your software and your connectivity are unimpaired. If something changes, such that you are dissatisfied, repeat 10) as though your software had changed. 12) If you really want to know how far you get, post to news.admin with an unlimited distribution and espouse silly ideas. 13) If you want even more response, post a chain letter. Ask for money. Or for post cards. Do it in alt.flame or talk.bizarre. Ask for a cannonical list of COBOL jokes. Do it in alt.folklore.computers. Crosspost to talk.bizarre. 14) For still more response, irritate someone with an impaired restraint gland so that they'll forge a sendsys that points to you. Or send out your own. Or rethink 0) :-). Good luck. -- Mike Murphy mrm@Sceard.COM ucsd!sceard!mrm +1 619 598 5874
mrm@sceard.Sceard.COM (M.R.Murphy) (05/31/91)
In article <7moR331w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: [...] > >No, it's like arguing that Postscript printers should report an error when >someone sends them an ASCII file, rather than just silently losing the whole >of the input and leading the user to wonder what exactly is going on. > >We have a printer here at work which silently junks any page containing >bitmap data sent in the wrong format. I think that said printer is profoundly >badly designed. You could read the manuals for the printer and find out how to download the short PostScript(tm) error routine that will barf out a page or so of error status when a job fails. The program can be modified as desired. Most folks don't like the waste of paper most of the time. Kind of like wasted bandwidth (like this posting, too:-) that would result from constant notification of the poster for a screwed up article. You can write short test files to see what is wrong. It is allowed. Or you can just stand in front of the printer with a quizzical look on your face and wait for the output. -- Mike Murphy mrm@Sceard.COM ucsd!sceard!mrm +1 619 598 5874
tp@mccall.com (Terry Poot) (05/31/91)
In article <1991May29.221015.1865@druid.uucp>, darcy@druid.uucp (D'Arcy J.M. Cain) writes: >that it conforms to the relevant specifications. No one on the "CNEWS >IS RIGHT" side suggests otherwise. It's the "CNEWS MUST BE WHINED AT >AND DIE" group that thinks that the posting software can slough that >stuff onto software running 6 or 7 hops away. How about the "WHY DIDN'T SOMEONE TELL US IN ADVANCE" group? Answers to that one that I've seen (all paraphrased), and my responses: 1) Because we didn't have time. This patch had to go out in such a hurry? 2) Nobody would fix the software anyway until we dropped articles Pure BS. I know I would have. I suspect the author of dxrn would have. Unfortunately, he was gone so I had to, and in a big hurry, causing me to screw up twice trying to get it out the door. 3) Who cares about the rest of the world as long as C news works. This seems to be the majority opinion (though not that of the more responsible people, who fall into groups 1 and 2). 4) We could have done a better job, we're sorry, and will try to avoid similar debacles in the future. OOPS! What was I thinking? That response has been totally absent. I must have dreamed it. >That's like arguing >that print spoolers should correct spelling because users make >spelling mistakes. How about arguing that before you switch the printer from ASCII to EBCDIC, you should tell the users that will be trying to print things. >So, the $64K question: Who or what should take responsibility for >the correctness of news headers? > A) The novice poster that probably hasn't even heard of an RFC. > B) The software that injects the message into the net. > C) The software that transports the message around the world. > >CNews suggests that the right answer is B. It's detractors claim that >C is right but they seem to argue that the only other alternative is A. The right answer is B. The software is correct. The transition was severly bungled. This is a human issue, not a software issue. C news switched from the de facto standard (B news Date: header processing) to the de jure standard (RFC822 date header processing) without sufficient warning to the non-C news community. (The suggestion that everyone should read news.software.b because so much of the net uses B and C news is utterly ridiculous.) I fully agree that RFC822 processing is correct and desirable. If the next release of nntp required message-id's and dates as I've been told the nntp spec requires, thus breaking most unix news readers, and this were done with little or no advance warning, there would be massive bitching. In my opinion, this would be justified. It seems the reaction of the "C news is holy" camp would probably be one of the first 3 above. -- Terry Poot <tp@mccall.com> The McCall Pattern Company (uucp: ...!rutgers!ksuvax1!deimos!mccall!tp) 615 McCall Road (800)255-2762, in KS (913)776-4041 Manhattan, KS 66502, USA
mathew@mantis.co.uk (Giving C News a *HUG*) (05/31/91)
louie@sayshell.umd.edu (Louis A. Mamakos) writes: > In article <7moR331w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) > >I fixed the broken software I was using as soon as I found out about the > >error. I would expect anyone else to do the same. THE POINT is that I didn't > >find out about the error for weeks, because C News made no attempt to tell m > >that I had failed its unnecessarily stringent requirements. > > Just two pointa after following this wonderful discussion: > > * What the #$^&* do you want for free? Just what does anyone pay for > cnews that justifies demands upon the authors to change fundmaental > design decisions on their part? I don't remember people *asking* for C News to silently delete articles. I suspect that it was one of those design decisions which seemed like a good idea at the time. If C News wishes to be pedantic about the headers, it should report its findings to someone who can do something about it. If it can't send its little pedantic reports to someone who can do something about the problem, then it shouldn't be pedantic. > * If you feel so strongly about the problems that cnews has, why not fix > it yourself, and offer to distribute your version to sites on the net that > are run by folks that agree with your point of view. I have already offered to spend my own spare time fixing the official C News for the next release. All I ask is access to a suitable UNIX machine in Cambridge. > I suppose that if you don't like the way this USENET is run [by > consensus, sort of], then go off and start your own. I'm quite happy with the way Usenet WAS run. C News seems to have changed the situation overnight. > I for one am tired > of hearing pot shots at positions and opinions that have been clearly > stated on many occasions. If they are clearly stated *wrong* positions, why not continue to take pot-shots at them? mathew
aipdc@castle.ed.ac.uk (Paul Crowley) (06/01/91)
In article <1991May31.002356.10376@druid.uucp> darcy@druid.uucp (D'Arcy J.M. Cain) writes: | 3 - The transport software should move news through a variety | of networks. To do this efficiently it should act on | a well defined set of rules. Messages that fail to | conform to these rules should not be passed. | |So if you agree with #1 and #2 and disagree with #3, on what |basis do you do so? Remember that the user is protected by |the posting software so buttinsky transport software is only |protecting bad posting software. I happen to believe that |if you are writing software that injects news into the news |system than the onus is on you to follow the standards and |not to expect Henry and Geoff to conform to yours. For what I think is the third time for me, and more if you count Mathew's contributions, there is no question where the "onus" lies - it is, clearly, with the posting software. No-one is asking Cnews to conform to standards I just made up. However, it still seems to me a necessary feature of quality software to behave sensibly given erronous input. I don't know what you mean by "protecting" bad software, since I still think that spiteful software is a bad idea. |I'll buy that. However I suspect we disagree on what the sensible |thing for transport software to do is. You seem to think it should |start a flood of reports back to the originator. I think that Cnews |does the sensible thing and drops them. There are two arguments here: one, is it desirable for Cnews to behave nicely given non-compliant articles to transport, and two, is it possible. Here you're arguing that it's not possible. I note that no-one has commented on Kent Paul Dolan's suggestion, and conclude that no-one really cares whether it's possible or not since it's not desirable. (By behave nicely, I mean making some attempt to either ignore, fix, or report errors.) ____ \/ o\ Paul Crowley aipdc@castle.ed.ac.uk \ / /\__/ Part straight. Part gay. All queer. \/ "I say we kill him and eat his brain." "That's not the solution to _every_ problem, you know!" -- Rudy Rucker
eggert@twinsun.com (Paul Eggert) (06/01/91)
The recent brouhaha over C News's date parser has overlooked a useful property of the current situation. It's now possible to post an article that targets only B News sites: just munge its ``Date:'' line to be something acceptable to B News but not C News. This gives a control over article distribution that can be more useful than what the ``Distribution:'' line gives. To take advantage of this opportunity, I've written a program `bnewsonly' that munges ``Date:'' lines to be acceptable to B News but not C News. A companion program `cnewsonly' targets C News sites. By careful header formatting, these programs can even send articles only to sites running particular software releases, e.g. just C News sites running version 7-Sep-1990 or later. This gives news transport authors a new way to target their bug reports and upgrade reminders -- a long-needed ability that would have helped to prevent the recent uproar. A Usenet administrator can install these programs inside the news transport mechanism itself: when given an optional regular expression argument, the programs munge only those articles whose subject lines match the argument. To demonstrate how this works, I'll install bnewsonly 'jokes being missed' locally, and then post the programs' source code in part 2 of this message. Please look at the change in header lines between the two parts of this message to see exactly how it works.
tp@mccall.com (Terry Poot) (06/01/91)
In article <1991May31.124723.12465@micrognosis.co.uk>, nreadwin@micrognosis.co.uk (Neil Readwin) writes: >So ... I still say that I would rather lose news than have old new >recirculated. Maybe I'm missing something, a I haven't run B news for several years, but doesn't B news manage to avoid both problems? As I understand it, it does parse the date, and ignores old news, without dropping news on the floor that has otherwise been acceptable to the net for many years. However, as I've said, I have no problem with the transition to official standards, just the way it was accomplished. I do firmly agree that all software that generates non-compliant headers should be fixed. I disagree that a reasonable and effective way to accomplish that is to simply ignore all news from sites that run such software, without bothering to tell them about it. On the other hand, I've said that I don't know how many times, and the responses are generally on another tangent, so I give up. Nobody is going to do anything about this problem. Sites will eventually find out their postings are going into a black hole. After much frustration and ranting and raving at not having been told, they will fix their software. In other words, the current situation and attendant flamage will be with us for a long time, because nobody in the C news camp finds this situation unacceptable, and, of course, nobody else can do much about it. I don't think that's an appropriate way to improve the net, but I guess I'm in the minority here. -- Terry Poot <tp@mccall.com> The McCall Pattern Company (uucp: ...!rutgers!ksuvax1!deimos!mccall!tp) 615 McCall Road (800)255-2762, in KS (913)776-4041 Manhattan, KS 66502, USA
darcy@druid.uucp (D'Arcy J.M. Cain) (06/01/91)
In article <10739@castle.ed.ac.uk> aipdc@castle.ed.ac.uk (Paul Crowley) writes: >There are two arguments here: one, is it desirable for Cnews to behave >nicely given non-compliant articles to transport, and two, is it >possible. Here you're arguing that it's not possible. Nope. Just that a workable solution hasn't been found yet and that there is no compelling reason for C News to take half measures just because there are people who want to write news posting software without reading the relevant specifications. > I note that >no-one has commented on Kent Paul Dolan's suggestion, and conclude that >no-one really cares whether it's possible or not since it's not >desirable. You say no one has commented on Kent's suggestion? Maybe you mean Henry hasn't accepted his suggestion uncritically. He did, however, respond in <1991May29.223011.25730@zoo.toronto.edu> >In article <1991May29.180201.5365@zorch.SF-Bay.ORG> xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) writes: >>Instead of mailing a warning back each time an >>article is dropped, each site should roll a random >>number generator, with a small chance of mailing a >>warning, and a large chance of dropping the article >>silently. > >How do you set the odds? Remember, news software sticks around for a long >time and is often maintained by relatively naive people. You can't count >on having the probability changed when appropriate. Especially given your >desire to have the probability different for each incoming path, this is >not a trivial problem. There were also other comments and followups to this thread such as this one from Phillip Evans in <1991May30.030612.8395@cynic.wimsey.bc.ca>: >In article <1991May29.223011.25730@zoo.toronto.edu> > henry@zoo.toronto.edu (Henry Spencer) writes: > [ above message repeated ] >Have the authors of the news software send some kind of control message >with the desired value. They will no doubt be attuned to the volume of >bad posts and responses. Then Neil Rickert made a suggestion in <1991May31.060933.27374@mp.cs.niu.edu> further in this thread following up another message which stemmed from Kent's original suggestion. In response to this Henry had this to say in message <1991May31.162219.21079@zoo.toronto.edu>: >Various ideas along these lines have been proposed recently. It would >certainly be better than mailing a response. Doing it as a control message >would be the obvious method, but unfortunately some of the major gatewaying >packages do not propagate control messages. (And no, gatewaying is not a >trivial issue. There is a *lot* of it, and it accounts for a significant >fraction of the total traffic. It also probably accounts for a lot of >the ill-formed articles.) I'm not yet happy with any of the schemes, but >it seems a promising direction. So exactly what do you mean by "no-one has commented on Kent Paul Dolan's suggestion?" Do you mean that Henry and Geoff, after years of planning and studying the news transport system, including the RFCs, coding, testing and discussions in various forums did not immediately implement Kent's idea into the next patch? Do you mean that the comments were not to your liking? You certainly can't mean that absolutely no one has posted any sort of response to Kent's suggestion even if that is exactly what you said. All the evidence that I have seen points to the fact that Henry and Geoff are putting out well thought out, well designed software and that they are willing to consider any new ideas if those presenting them have made at least some effort to investigate the effects of their ideas. They seem to be willing to engage in dialogue. They don't seem to be overly impressed by whining and foot stomping and I would be surprised if they considered someone changing their name to "CNEWS MUST DIE" to be an example of reasoned and logical debate. -- D'Arcy J.M. Cain (darcy@druid) | D'Arcy Cain Consulting | There's no government Toronto, Ontario, Canada | like no government! +1 416 424 2871 |
kevin@rotag.mi.org (Kevin Darcy) (06/02/91)
In article <T6mR328w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >rickert@mp.cs.niu.edu (Neil Rickert) writes: >> ...we are tired at your continued attempts to >> harrass the totally innocent designers of Cnews, and the totally innocent >> sysadmins who have made an honest and considered judgement to use Cnews. > >But C News is the program which is meting out the punishment! > >Or is it "only following orders", as the famous saying goes? Perhaps "taking out the garbage" is a more apt term. Or "separating the chaff from the wheat". Mathew: it is not the job of C News to de-louse every bad header it sees, nor to yak-yak to humans about the sad state of posting software. It's purpose is to transfer news to and from disks and networks as smoothly and efficiently as possible. C NEWS IS A NEWS *TRANSPORT*, NOT A NEWS *SCHOOLMARM*!! Given the obvious advantages and increasing popularity of C News over other news transports, and the fact that it follows the RFC in this regard, I consider it perfectly reasonable to expect sites with broken posting software to adapt to C News, rather than the other way around. Why should the tail wag the dog? ^^^^^^^^ ^^^^^^^^^^^---- Ob. Hackneyed Phrase >mathew - Kevin CNEWS MUST LIVE! CNEWS MUST FLOURISH!
brendan@cs.widener.edu (Brendan Kehoe) (06/02/91)
From: mathew@mantis.co.uk (Giving C News a *HUG*)
Newsgroups: news.admin,news.software.b
Date: 31 May 91 15:22:14 GMT
If C News wishes to be pedantic about the headers, it should report its
findings to someone who can do something about it. If it can't send its
little pedantic reports to someone who can do something about the problem,
then it shouldn't be pedantic.
It DOES, in the form of daily reports--the problem is that everyone "can"
do something about it, just not everyone *wants* to. If your upstream
site does want to, then problem solved; if he or she doesn't, oh well.
I'm quite happy with the way Usenet WAS run. C News seems to have
changed the situation overnight.
I'm sorry this crusade of yours has ruined it for you.
If they are clearly stated *wrong* positions, why not continue to take
pot-shots at them?
Wrong by whose standards? Yours? Pfft.
Brendan
--
Brendan Kehoe - Widener Sun Network Manager - brendan@cs.widener.edu
Widener University in Chester, PA A Bloody Sun-Dec War Zone
"For a healthy environment, today and tomorrow . . .
this is a biodegradable plastic message."aipdc@castle.ed.ac.uk (Paul Crowley) (06/02/91)
In article <1991Jun1.152224.1293@druid.uucp> darcy@druid.uucp (D'Arcy J.M. Cain) writes: > >Nope. Just that a workable solution hasn't been found yet and that there >is no compelling reason for C News to take half measures just because >there are people who want to write news posting software without reading >the relevant specifications. Strange, most of the real world has faced up to the fact that software contains errors, and good software takes account of the fact. >> I note that >>no-one has commented on Kent Paul Dolan's suggestion, and conclude that >>no-one really cares whether it's possible or not since it's not >>desirable. [list of responses deleted] >So exactly what do you mean by "no-one has commented on Kent Paul Dolan's >suggestion?" My apologies. At the time I wrote the article, there had been only one post pointing out problems with his approach. That post was in news.admin rather than news.software.b, and so I didn't see it. I have now started following both groups, and there seems to be a decent debate on the subject. And yes, it does seem that Henry and Geoff are trying to find ways around the problem. What gets me is that they're the only people who seem to recognise that it's a good idea to try to solve it. Thankfully, Henry and Geoff are the people writing the software, so should any of the approaches in the air at the moment be turned to viable solutions there's a good chance they'll be used. ____ \/ o\ Paul Crowley aipdc@castle.ed.ac.uk \ / /\__/ Part straight. Part gay. All queer. \/ "I say we kill him and eat his brain." "That's not the solution to _every_ problem, you know!" -- Rudy Rucker
csu@alembic.acs.com (Dave Mack) (06/03/91)
In article <RVHT35w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >louie@sayshell.umd.edu (Louis A. Mamakos) writes: >> I for one am tired >> of hearing pot shots at positions and opinions that have been clearly >> stated on many occasions. > >If they are clearly stated *wrong* positions, why not continue to take >pot-shots at them? There is a newsgroup for this sort of thing. It's called alt.sadonecrobestiality. It was created for people just like you, who won't stop beating a dead horse. Followups to alt.sadonecrobestiality. -- Dave Mack
jensen@login.dkuug.dk (j e n s e n) (06/03/91)
eggert@twinsun.com (Paul Eggert) writes: >[...] >A Usenet administrator can install these programs inside the news >transport mechanism itself: when given an optional regular expression >argument, the programs munge only those articles whose subject lines >match the argument. To demonstrate how this works, I'll install > bnewsonly 'jokes being missed' >locally, and then post the programs' source code in part 2 of this >message. Please look at the change in header lines between the two >parts of this message to see exactly how it works. You didn't post that second part did you ?-) -- At least I didn't receive it, and I'm not exactly surrounded by Cnews-sites - as far as I know. I received part one thru the following path: Path: dkuug!sunic!hagbard!eru!bloom-beacon!ora!camb.com!spdcc \ !mintaka!think.com!zaphod.mps.ohio-state.edu!usc!ucla-cs!twinsun!eggert == -- . . . . . . . . . . . . . . . . . . . . . . . . . . j e n s e n
mathew@mantis.co.uk (Giving C News a *HUG*) (06/03/91)
darcy@druid.uucp (D'Arcy J.M. Cain) writes: > 1 - Anyone who posts to the net should make a reasonable effort > to get the headers right. [...] > 2 - The posting software should not try to decide whether the > subject reflects the content or whether the cross-posting > is relevant. It should however make sure that every header > line follows the RFCs. If not it should at least notify > the user immediately that there is a problem. It might > also try to correct the situation with or without the > user's help depending on the nature of the problem. > > 3 - The transport software should move news through a variety > of networks. To do this efficiently it should act on > a well defined set of rules. Messages that fail to > conform to these rules should not be passed. > I'll agree with all three of those statements. Where we differ is that I have a fourth statement: 4 - When something goes wrong, someone who can fix the problem should be told about it. If C News wishes to drop messages which are not RFC822 conformant then I'll agree to that, SO LONG AS it reports that it has dropped them. Contrariwise, if it is unable to send reports, then it must do its utmost not to drop articles. > Remember that the user is protected by > the posting software so buttinsky transport software is only > protecting bad posting software. No, it's telling the person who is using the bad posting software that there is something wrong with his software. If he isn't told, he won't know to fix it, and might carry on posting for months having his articles dropped silently. > I happen to believe that > if you are writing software that injects news into the news > system than the onus is on you to follow the standards and > not to expect Henry and Geoff to conform to yours. Me too. However I also believe that since we KNOW that there will always be imperfect software out there injecting articles into the news stream, it makes sense to have C News behave in a safe and helpful way when it encounters bad articles. It certainly doesn't make sense to have C News ASSUME that all posting software is perfect, since this assumption is demonstrably false. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/03/91)
darcy@druid.uucp (D'Arcy J.M. Cain) writes: > In article <T6mR328w164w@mantis.co.uk> CNEWS MUST DIE! writes: > >But C News is the program which is meting out the punishment! > >Or is it "only following orders", as the famous saying goes? > > What orders? The Holy RFCs. > You seem to be the one with the orders and Cnews > certainly isn't following them. My statements as to what C News should do are recommendations based on numerous software engineering principles and social considerations. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/03/91)
tp@mccall.com (Terry Poot) writes: > However, as I've said, I have no problem with the transition to official > standards, just the way it was accomplished. I do firmly agree that all > software that generates non-compliant headers should be fixed. I disagree > that a reasonable and effective way to accomplish that is to simply ignore > all news from sites that run such software, without bothering to tell them > about it. This is my position as well. > On the other hand, I've said that I don't know how many times, and the > responses are generally on another tangent, so I give up. I have no plans to give up, because (a) I believe that C News's current behaviour is damaging to the people who use Usenet, and (b) I'm exceptionally stubborn when fighting things which I genuinely believe to be wrong. By the latter I mean that those who want me to shut up about this issue will have to convince me that C News's behaviour is correct; I won't just give up. mathew [ You have been warned :-) ]
mathew@mantis.co.uk (Giving C News a *HUG*) (06/03/91)
mrm@sceard.Sceard.COM (M.R.Murphy) writes: > mathew, this is the way you do it. > > 0) You decide you want to use USENET. > 1) You install News software. [...and so on...] > 5) You repeat 4) until you are satisfied that your postings are not pond > scum [...] > 11) If you want to know how far you get, post to misc.test with an unlimited > distribution. You are allowed to do this with reasonable frequency to > see if your software and your connectivity are unimpaired. If something > changes, such that you are dissatisfied, repeat 10) as though your > software had changed. As I have mentioned several times, I did all of this. The software I was using was perfectly OK according to every test I was able to do from here. I didn't get any error reports, I didn't notice anything untoward. Postings to misc.test using the 'bad' software worked -- I got responses. I therefore assumed everything was OK -- until suddenly our feed site changed to the latest C News, and I got a mail message from the administrator there saying "Oh, all your postings from <blah> are being dropped, here are the log entries". Since this was at least a number of weeks after the changes to C News were released, I am forced to assume that for those weeks all my articles were being dropped by all the sites running the latest C News, and therefore not being propogated properly. I'll admit that I didn't initially spend much time studying the headers produced by the software I was using. I didn't spend hours checking the syntax with that in the RFCs. Who does? Yes, initially I overlooked the difference between "Wed, 19 Oct 91" and "Wed 19 Oct 91". Should that be a capital offense? Later, when I began to suspect that things were amiss, I asked politely in news.software.b how C News treated bad date headers. Nobody even bothered to reply. Then, as I say, I got some mail saying that all my articles had been dropped for the past few months. > 14) For still more response, irritate someone with an impaired restraint gla > so that they'll forge a sendsys that points to you. It has been pointed out before, but I'll point it out again just in case: In order to overload this site, you would have to send a sufficiently large quantity of junk that you would certainly overload ibmpcug, and probably overload UKC -- hence screwing up Usenet access for much of the UK. Such net.terrorism would, I feel, be inadvisable. mathew
cudep@warwick.ac.uk (Ian Dickinson) (06/03/91)
In article <BRENDAN.91Jun1191245@shirley.cs.widener.edu> brendan@cs.widener.edu writes: >mathew writes: > If C News wishes to be pedantic about the headers, it should report its > findings to someone who can do something about it. If it can't send its > little pedantic reports to someone who can do something about the problem, > then it shouldn't be pedantic. >It DOES, in the form of daily reports--the problem is that everyone "can" >do something about it, just not everyone *wants* to. If your upstream >site does want to, then problem solved; if he or she doesn't, oh well. Unfortunately, at the moment Cnews doesn't keep a copy of the article hanging around for a newsadmin to look at. Since I can't currently talk nntp, I can't look at the article text and warn the relevant newsadmin/author. Yours hoping-this'll-be-fixed-in-the-next-patch-edly, -- \/ato /'\ /`\ Ian Dickinson TED KALDIS FOR PRESIDENT! /^^^\/^^^\ vato@warwick.ac.uk /TWIN/TEATS\ @c=GB@o=University of Warwick@ou=Computing Services@cn=Ian Dickinson / \
mathew@mantis.co.uk (Giving C News a *HUG*) (06/03/91)
darcy@druid.uucp (D'Arcy J.M. Cain) writes: > I would be surprised if they > considered someone changing their name to "CNEWS MUST DIE" to be an > example of reasoned and logical debate. I admit that I made a mistake there; I severely underestimated the number of humour-impaired people reading this newsgroup. I hope that the new comment is more to your liking. mathew [ There doesn't seem much point in having the name field contain my real name, since it's there anyway as the login name. ]
mathew@mantis.co.uk (Giving C News a *HUG*) (06/03/91)
brendan@cs.widener.edu (Brendan Kehoe) writes: > From: mathew@mantis.co.uk (Giving C News a *HUG*) > If C News wishes to be pedantic about the headers, it should report its > findings to someone who can do something about it. If it can't send its > little pedantic reports to someone who can do something about the problem, > then it shouldn't be pedantic. > > It DOES, in the form of daily reports No, it DOES NOT. The daily reports are not sent to someone who can do something about the problem; they are sent to someone who can ask someone else to do something about the problem. There is an important difference, which we can sum up by observing that the larger the number of humans involved in propogating a message, the less reliable the propogation is. > --the problem is that everyone "can" > do something about it, just not everyone *wants* to. If your upstream > site does want to, then problem solved; if he or she doesn't, oh well. Right. And I'm saying that it is wrong to make correct operation of the net depend unnecessarily on every system administrator being competent and having the good will to send off warnings by mail. > If they are clearly stated *wrong* positions, why not continue to take > pot-shots at them? > > Wrong by whose standards? Yours? General software engineering principles. mathew
herrickd@iccgcc.decnet.ab.com (06/04/91)
In article <1991May29.224559.26237@zoo.toronto.edu>, henry@zoo.toronto.edu (Henry Spencer) writes: > In article <1991May28.213321.17039@skypod.guild.org> scott@skypod.guild.org (Scott Campbell) writes: >>>There may not be one "first C News site" to see the article; there may be >>>dozens or even hundreds of them, if the article first passed through a >>>B News site with a large news fanout. >> >>...like uunet? > > That's the example I had in mind, although it's now obsolete since uunet > is running C News now. I seem to remember reading a quiet comment about inability to persuade uunet to run C News in some group a very few months ago. Now it has quietly changed. Then there was a single posted, rather vehement, complaint about BITFTP originating too much traffic. And it quietly stopped. Amazing how effective the well reasoned complaints are. dan herrick herrickd@iccgcc.decnet.ab.com
rickert@mp.cs.niu.edu (Neil Rickert) (06/04/91)
In article <qX5y318w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >brendan@cs.widener.edu (Brendan Kehoe) writes: >> From: mathew@mantis.co.uk (Giving C News a *HUG*) >> If C News wishes to be pedantic about the headers, it should report its >> findings to someone who can do something about it. If it can't send its >> little pedantic reports to someone who can do something about the problem, >> then it shouldn't be pedantic. >> >> It DOES, in the form of daily reports > >No, it DOES NOT. The daily reports are not sent to someone who can do >something about the problem; they are sent to someone who can ask someone >else to do something about the problem. There is an important difference, >which we can sum up by observing that the larger the number of humans >involved in propogating a message, the less reliable the propogation is. Your mistake, not Brendan's. The trouble is, you are acting very much as an injured party, just because your articles were dropped by Cnews at another site. You were not an injured party. No other site has any obligation whatsoever to carry your news. When a site discards an article, the only injured parties are those who read news at that site, and the sites that are directly fed. There are several sites around which silently discard all articles posted to 'alt.sex.pictures'. Do they have an obligation to send email to the authors of those articles? I don't carry any 'alt' at my site. Must I notify all authors of articles in alt.*? Cnews correctly notifies the responsible party - namely the news admin at the site where the article is dropped. Many of us would like to go a step further and try to get word back to the originating site. But we recognize that this is very difficult to do in a safe but effective manner. The discussion on this topic before you joined in already revealed the concern. What we are all reacting to is your attitude of acting as if everyone is picking on you. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
sloane@kuhub.cc.ukans.edu (06/04/91)
In article <Vu3y312w164w@mantis.co.uk>, mathew@mantis.co.uk (Giving C News a *HUG*) writes: > By the latter I mean that those who want me to shut up about this issue will > have to convince me that C News's behaviour is correct; I won't just give up. If you want to see what happens when a message is sent to the author of an invalid message, try posting an invalid control message, for example: Control: Cancel <1@site.uucp> I once did this by accident (my email gateway dumped some bounce messages to the net, and I thought I would be nice and clean them up). It turns out that notes sends email to the originator of the invalid control message. I couldn't get much done for several days due to all the email I got. -- USmail: Bob Sloane, University of Kansas Computer Center, Lawrence, KS, 66045 E-mail: sloane@kuhub.cc.ukans.edu, sloane@ukanvax.bitnet, AT&T: (913)864-0444
dan@kfw.COM (Dan Mick) (06/04/91)
In article <Vw2y39w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes:
<more sanctimonious bs>
Okay, here's a response that's about as much use as your postings on the
subject of C News and its cancellation behaviour have been:
Shut up. Just shut up. SHUT UP!!!!jbrown@locus.com (Jordan Brown) (06/04/91)
In article <1991May31.101334@mccall.com> tp@mccall.com (Terry Poot) writes: >(The suggestion that everyone should read news.software.b because so >much of the net uses B and C news is utterly ridiculous.) News.software.b is the de facto newsgroup for discussing generic news standards and other technical matters. I would consider it at least a little foolish for anybody involved heavily in developing a news-related package NOT to read news.software.b. I don't know of any more appropriate place to reach news developers. This is not to say that I believe that it's an ideal name, just that it *is* where the technical discussions occur. I'd prefer (for instance) news.software.transport, news.software.standards, etc.
cudep@warwick.ac.uk (Ian Dickinson) (06/04/91)
In article <Vu3y312w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >By the latter I mean that those who want me to shut up about this issue will >have to convince me that C News's behaviour is correct; I won't just give up. Nope - those that want you to shut up will just put you in their kill files :-) -- \/ato /'\ /`\ Ian Dickinson TED KALDIS FOR PRESIDENT! /^^^\/^^^\ vato@warwick.ac.uk /TWIN/TEATS\ @c=GB@o=University of Warwick@ou=Computing Services@cn=Ian Dickinson / \
jmaynard@thesis1.med.uth.tmc.edu (Jay Maynard) (06/04/91)
In article <Vu3y312w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >By the latter I mean that those who want me to shut up about this issue will >have to convince me that C News's behaviour is correct; I won't just give up. *plonk* I have no reason to believe that you'll ever be convinced of anything of the sort. Since you're patently wrong, and will insist on being loudly, publicly, vociferously wrong, there's a new entry in my global kill file: /From: mathew@mantis.co.uk/h:j -- Jay Maynard, EMT-P, K5ZC, PP-ASEL | Never ascribe to malice that which can jmaynard@thesis1.med.uth.tmc.edu | adequately be explained by stupidity. "As far as I'm concerned, 'average skills' means you know that the pointy end of the drill makes the holes." -- Ron Wanttaja, _Kitplane Construction_
mathew@mantis.co.uk (Giving C News a *HUG*) (06/04/91)
rickert@mp.cs.niu.edu (Neil Rickert) writes: > When a site discards an article, the only injured parties are those who > read news at that site, and the sites that are directly fed. All sites receive at least some of their news via C News sites. Therefore all sites are forced to put up with C News's deletion of articles whether they like it or not. > There are > several sites around which silently discard all articles posted to > 'alt.sex.pictures'. Indeed. However, anyone who wants to get a feed of alt.sex.pictures can do so. I CANNOT get a complete newsfeed which includes the C News-deleted articles even if I want to. That's the crucial difference. There are other differences, too; for example, the readers of Usenet know if an entire newsgroup is being chopped. They don't know if random articles are being chopped. This is why people like John 'DDMI' Palmer, who apparently provide filtered news to their clients without telling them that it is being filtered, end up getting flamed. > Cnews correctly notifies the responsible party - namely the news admin at > the site where the article is dropped. So you admit that the admins running C News are the ones responsible for articles being lost? That's very decent of you. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/04/91)
sloane@kuhub.cc.ukans.edu writes: > If you want to see what happens when a message is sent to the author > of an invalid message, try posting an invalid control message [...] This is Classic Objection #1: "You can't have every site send a mail message to the author of the bad article!" Classic Response #1: "Very good. If you actually read through this discussion, you'll note that nobody is suggesting that every site send a mail message to the author of the bad article." > It turns out that notes sends email to the originator of the invalid > control message. I couldn't get much done for several days due to all > the email I got. Notes is broken, then. Let's fix it. mathew
csu@alembic.acs.com (Dave Mack) (06/05/91)
In article <imP1313w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >rickert@mp.cs.niu.edu (Neil Rickert) writes: >> Cnews correctly notifies the responsible party - namely the news admin at >> the site where the article is dropped. > >So you admit that the admins running C News are the ones responsible for >articles being lost? That's very decent of you. Well of course we're the ones responsible. We installed the March 91 Cnews patches *knowing* that it would mean that some articles would be discarded. So what? Here's a clue for you, mathew: we are not obligated to ensure that *any* article gets propagated. And if any article doesn't conform to the relevant RFCs, there are damned good reasons for *not* propagating it, as has already been explained to you in excruciating detail. If the software you use to post articles with doesn't produce valid headers, that is *your* problem, not ours. Bitch to the people responsible for your posting software. I agree that it would be nice (though certainly not mandatory) to warn people whose software is screwed up that it is, but only if the overall impact on the net is negligible. If Henry and Geoff come up with a patch that does this is some sensible fashion, I'll install it. Until then, shit happens. -- Dave Mack
brad@looking.on.ca (Brad Templeton) (06/05/91)
Strictly, nobody is "responsible" for losing articles, since nobody has a responsibility to propagate articles. You start thinking that other people have a duty to send your articles (particularly your illegal-format articles) around the world and you're getting some very wrong ideas. Usenet is fairly anarchic, however, the one thing that must be agreed upon is the format of the articles. Indeed, the format of the articles, along with things implied by this format (like newsgroups) is actually what defines USENET -- USENET is sites that pass around articles in the specified formats. It would be nice if you were told, but as the implementors did not see an easy way to do this at the time, the only answer is tough titty. Perhaps at a later time this will be done. -- Brad Templeton, ClariNet Communications Corp. -- Waterloo, Ontario 519/884-7473
scott@skypod.guild.org (Scott Campbell) (06/05/91)
In article <Vu3y312w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >I have no plans to give up, because (a) I believe that C News's current >behaviour is damaging to the people who use Usenet, and (b) I'm exceptionally >stubborn when fighting things which I genuinely believe to be wrong. > >By the latter I mean that those who want me to shut up about this issue will >have to convince me that C News's behaviour is correct; I won't just give up. I don't know whether CNEWS' current behaviour is "correct" or not, but I have noticed that since many sites near here upgraded, I no longer have to run expire manually 15 times a day to get rid of megabytes of month old news that is coming back to haunt me... maybe it was just a coincidence... scott -- Scott J.M. Campbell scott@skypod.guild.org Skypod Communications Inc. ..!gatech!dscatl!daysinns!skypod!scott 1001 Bay Street, Suite 1210 ..!uunet!utai!lsuc!becker!skypod!scott Toronto, Ont. (416) 924-4059 ..!epas.utoronto.ca!nyama!skypod!scott
mathew@mantis.co.uk (Giving C News a *HUG*) (06/05/91)
csu@alembic.acs.com (Dave Mack) writes: > In article <imP1313w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * > >So you admit that the admins running C News are the ones responsible for > >articles being lost? That's very decent of you. > > Well of course we're the ones responsible. We installed the March 91 > Cnews patches *knowing* that it would mean that some articles would > be discarded. So what? > > Here's a clue for you, mathew: we are not obligated to ensure that > *any* article gets propagated. Well, no. But if you feed any other site, that site is probably under the impression that it is getting a feed of all the articles in the newsgroups it requests. I do hope you will explain that you are, in fact, deliberately and unnecessarily silently discarding articles whether they like it or not. Furthermore, if your attitude to people at other sites making errors is "fuck you", then presumably you don't send bounce messages for badly-addressed mail. > And if any article doesn't conform > to the relevant RFCs, there are damned good reasons for *not* > propagating it, as has already been explained to you in excruciating > detail. Indeed. If you want to throw away any article which is even slightly wrong, then that's entirely up to you. Just make sure you tell the people who think that you are forwarding articles. To turn your argument around: If a mail message is not correctly addressed, there are damn good reasons for *not* forwarding it, as has already been explained in excruciating detail. Therefore we should silently delete badly-addressed mail messages. Eventually people sending mail will correct their software and learn to be more careful so that erroneous messages do not occur. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/05/91)
brad@looking.on.ca (Brad Templeton) writes: > Strictly, nobody is "responsible" for losing articles, since nobody has > a responsibility to propagate articles. You start thinking that other people > have a duty to send your articles (particularly your illegal-format articles) > around the world and you're getting some very wrong ideas. [...] > It would be nice if you were told, but as the implementors did not see an > easy way to do this at the time, the only answer is tough titty. The problem is that this answer detracts from the usefulness of Usenet. OK, you say. One of the conditions of using Usenet is that it is possible that everything you type will be thrown away by some piece of software somewhere, which will make no attempt whatsoever to inform you that anything is wrong. Now, how many people would really use such a system? How many people would take the time to sit and correct mistakes, to research sources, to carefully edit considered responses, if all that work could end up being thrown away through no fault of the user, and without the user being told? I know I wouldn't. And before some smart-arse comments that I don't anyway, I shall remark that in certain other newsgroups I *do* spend a lot of time on postings, I *do* carefully edit considered responses, and I *do* research sources. If you think my postings in news.software.b are worthless, that's your opinion; my impression is that my postings in some other newsgroups are not considered worthless. Do we *want* Usenet to be a system where users are *discouraged* from spending time on their postings? If not, then we must try to make sure that posting an article carries a reasonably high probability that either the article will be propogated as expected or that the user will get to hear about it (perhaps via *his* system administrator). mathew
jhenders@wimsey.bc.ca (John Henders) (06/06/91)
Matthew writes: >..... If you think my postings in news.software.b are worthless, that's >your opinion; my impression is that my postings in some other newsgroups are >not considered worthless. > I think Matthew's postings here are quite valuable, as so far they are the only way I could have found out that my software is broken and get it fixed. Perhaps this is the intended mechanism to report errors in the new C news. The first person to find out about a problem starts a huge flame war, thus notifying others of the problem. Seems to work, thought it's not as efficient as some might like. John Henders Vancouver,B.C.
rickert@mp.cs.niu.edu (Neil Rickert) (06/06/91)
In article <m6P3324w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >If a mail message is not correctly addressed, there are damn good reasons for >*not* forwarding it, as has already been explained in excruciating detail. >Therefore we should silently delete badly-addressed mail messages. Eventually >people sending mail will correct their software and learn to be more careful >so that erroneous messages do not occur. Mail is very different from news. I strongly support Henry and Geoff in maintaining high standards for new. For mail directed to a mailing list high standards are also needed, because large mailing lists have many of the characteristics of news. But those who have suffered my comments in other forums know I will fight to avoid setting the same high standards for general email. Often email is used as a basic communication method to get help for solving problems, so a more tolerant attitude is required. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
aipdc@castle.ed.ac.uk (Paul Crowley) (06/06/91)
In article <F4q3327w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >Do we *want* Usenet to be a system where users are *discouraged* from >spending time on their postings? It seems odd to try and defend the need for USENET _in_ USENET. If you're here, you surely see the need for reliably posted articles. Supposing a moderator got cut off in this manner? Supposing the local group guru were accidentally shut out (for example, der Mouse in comp.windows.x). Supposing that a discussion is going on that could lead to an important improvement in a piece of software, or check the onslaught of software patents, or some such? The unique ability of news to solve such problems would be seriously impaired. Why do you think USENET is useful? You can't have it both ways. ____ \/ o\ Paul Crowley aipdc@castle.ed.ac.uk \ / /\__/ Part straight. Part gay. All queer. \/ "I say we kill him and eat his brain." "That's not the solution to _every_ problem, you know!" -- Rudy Rucker
herrickd@iccgcc.decnet.ab.com (06/06/91)
In article <1991May31.101334@mccall.com>, tp@mccall.com (Terry Poot) writes: > > How about the "WHY DIDN'T SOMEONE TELL US IN ADVANCE" group? Answers to > that one that I've seen (all paraphrased), and my responses: > > 1) Because we didn't have time. > > This patch had to go out in such a hurry? > > 2) Nobody would fix the software anyway until we dropped articles > > Pure BS. I know I would have. I suspect the author of dxrn would have. > Unfortunately, he was gone so I had to, and in a big hurry, causing me > to screw up twice trying to get it out the door. > > 3) Who cares about the rest of the world as long as C news works. > > This seems to be the majority opinion (though not that of the more > responsible people, who fall into groups 1 and 2). > > 4) We could have done a better job, we're sorry, and will try to avoid > similar debacles in the future. > > OOPS! What was I thinking? That response has been totally absent. I > must have dreamed it. > You left out the response by the people who paid attention. Their response was "It was announced about three months in advance." The phone companies have a similar problem when they split an area code. They plaster billboards and the sides of buses and the newspapers and pay phones with announcements that the area code is going to change. There will be a period when you can dial either the old way or the new way and your call will go through. Then there will be a day when the old way will not work. When that fateful day comes around, some people continue on about their business without noticing. Others scream bloody murder because THEY broke my system and nobody warned me. The discussions of this on comp.dcom.telecom have been quite amusing. So have the parallel discussions here. dan herrick herrickd@iccgcc.decnet.ab.com
herrickd@iccgcc.decnet.ab.com (06/06/91)
In article <1991May31.145603.10105@sceard.Sceard.COM>,
mrm@sceard.Sceard.COM (M.R.Murphy) made an important contribution
for us beginners.
Chip or Spaf, would you capture his advice into the periodic
postings?
Please?
dan herrickrandy@m2xenix.psg.com (Randy Bush) (06/07/91)
mathew@mantis.co.uk (Giving C News a *HUG*) writes: > Well, no. But if you feed any other site, that site is probably under the > impression that it is getting a feed of all the articles in the newsgroups it > requests. I do hope you will explain that you are, in fact, deliberately and > unnecessarily silently discarding articles whether they like it or not. Nope. Deleting garbage that is trying to pretend it is an article but not succeeding. > Furthermore, if your attitude to people at other sites making errors is "fuck > you" Nah. Just the ones who decide to defend their errors to the death. -- randy@psg.com ..!uunet!m2xenix!randy
fenner@jazz.psu.edu (Bill Fenner) (06/07/91)
In article <m6P3324w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: |csu@alembic.acs.com (Dave Mack) writes: |> Here's a clue for you, mathew: we are not obligated to ensure that |> *any* article gets propagated. | |To turn your argument around: | |If a mail message is not correctly addressed, there are damn good reasons for |*not* forwarding it, as has already been explained in excruciating detail. |Therefore we should silently delete badly-addressed mail messages. Eventually |people sending mail will correct their software and learn to be more careful |so that erroneous messages do not occur. The mail analogy is flawed, though, Mathew. There is only ever one copy of an E-mail message, and it only ever has one destination. Therefore, there's no ambiguity about whether bouncing it will create huge amounts of mail to the sender; you know you have the only copy. (Erik Fair, and mailing lists in general, excluded. :-) In USENET, you have one of many copies of an article, and it doesn't have any specific destionation (hint: "world" is not a specific enough destination.) . There is no easy solution to the news problem. A control message with a modified Message-ID is the best solution I've seen so far. Bill -- Bill Fenner fenner@jazz.psu.edu ..psuvax1!hogbbs!wcfpc!wcf wcf@hogbbs.scol.pa.us (+1 814 238-9633 2400MNP5)
sarima@tdatirv.UUCP (Stanley Friesen) (06/07/91)
In article <1991Jun5.113003.25300@skypod.guild.org> scott@skypod.guild.org (Scott Campbell) writes:
<I don't know whether CNEWS' current behaviour is "correct" or not, but I have
<noticed that since many sites near here upgraded, I no longer have to run
<expire manually 15 times a day to get rid of megabytes of month old news
<that is coming back to haunt me... maybe it was just a coincidence...
You mean *that* was the source of my #*!*!!# i-node leak!!! Geez!
All I can say is - why wasn't this fixed a long time ago!?
Good job C-news.
--
---------------
uunet!tdatirv!sarima (Stanley Friesen)xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) (06/07/91)
rickert@mp.cs.niu.edu (Neil Rickert) writes: > Mail is very different from news. [...] > But those who have suffered my comments in other > forums know I will fight to avoid setting the same > high standards for general email. Often email is > used as a basic communication method to get help > for solving problems, so a more tolerant attitude > is required. An utterly bogus distinction; go to comp.theory, rec.games.programmer, comp.sys.amiga.programmer, alt.graphics.pixutils, comp.fonts, or any of hundreds of other newsgroups, and you will see, over time, _thousands_ of articles starting off "Can anyone tell me how to..." Often Usenet news is used as a basic communication method to get help for solving problems, so a more tolerant attitude is required. Making functional software at sites different from the one where your changes are installed suddenly non-functional is simply unforgivable professional behavior. End of discussion. Kent, the man from xanth. <xanthian@Zorch.SF-Bay.ORG> <xanthian@well.sf.ca.us>
mathew@mantis.co.uk (Giving C News a *HUG*) (06/07/91)
scott@skypod.guild.org (Scott Campbell) writes: > I don't know whether CNEWS' current behaviour is "correct" or not, but I have > noticed that since many sites near here upgraded, I no longer have to run > expire manually 15 times a day to get rid of megabytes of month old news > that is coming back to haunt me... maybe it was just a coincidence... Well, it's certainly nothing to do with whether C News reports errors or not, which is what we're talking about. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/07/91)
rickert@mp.cs.niu.edu (Neil Rickert) writes: > In article <m6P3324w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * > >If a mail message is not correctly addressed, there are damn good reasons fo > >*not* forwarding it, as has already been explained in excruciating detail. > >Therefore we should silently delete badly-addressed mail messages. Eventuall > >people sending mail will correct their software and learn to be more careful > >so that erroneous messages do not occur. > > Mail is very different from news. So tell me: why is it OK to silently throw away news a person has written, but not OK to silently throw away mail a person has written? > I strongly support Henry and Geoff in maintaining high standards for news. So do I. All that I ask is that they report errors. mathew
tp@mccall.com (Terry Poot) (06/07/91)
In article <4812.284e0cf6@iccgcc.decnet.ab.com>, herrickd@iccgcc.decnet.ab.com writes: >You left out the response by the people who paid attention. Their >response was "It was announced about three months in advance." Where? With what subject line? >The phone companies have a similar problem when they split an area >code. They plaster billboards and the sides of buses and the newspapers The usenet equivalent of plastering it on a billboard would be posting it to news.announce.important, which has not been done, and would still be in order. I do recall seeing some discussion about C news' handling of articles with bogus headers. Given that I was running a news system that was generating articles accepted world wide, including by the C news system that is my upstream feed (and thankfully has not yet installed the patch), I didn't pay much attention to it. My software didn't generate bogus headers, it generated headers acceptable to ANU News, C news, and B news, at the very least. Now, had there been a discussion about C news dropping articles with valid headers, or previously valid headers, or wording of that nature in the subject line, in news.admin, I would have seen it and acted on it. Do you read absolutely every article in this group? I don't. I read the articles that I'm interested in, and those that seem to have a bearing on maintaining the functioning of my news system. I only wandered into this thread by mistake. If I see a thread about C news, I skip it because I don't run C news. Take the lists of dropped messages being posted in news.lists as an example. 'C news non-header line report'? Why would I read that. At best, it looks like a bug report about C news posted in the wrong group. If you are trying to warn people, it makes sense to post in a forum read by those people, with a subject line designed to attract their attention. Subject: ATTENTION non-C news newsadmins, important warning would certainly have gotten my attention, posted to news.admin, news.announce.important, or news.software.anu-news. I don't read news.software.b because I don't run B news or C news. I suspect you don't read news.software.anu-news for similar reasons. Posting something about C news dropping articles with bogus headers is about as much warning as the legal notices in the back of the newspaper that few people ever read. They aren't designed to notify anybody, just to be able to say "we did notify the public" at a later date. -- Terry Poot <tp@mccall.com> The McCall Pattern Company (uucp: ...!rutgers!ksuvax1!deimos!mccall!tp) 615 McCall Road (800)255-2762, in KS (913)776-4041 Manhattan, KS 66502, USA
rickert@mp.cs.niu.edu (Neil Rickert) (06/07/91)
In article <oT8633w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >rickert@mp.cs.niu.edu (Neil Rickert) writes: >> Mail is very different from news. > >So tell me: why is it OK to silently throw away news a person has written, >but not OK to silently throw away mail a person has written? Firstly C-news does not silently throw away news. It reports it to the local administrator. Secondly email is from one person to another. One of the two should be notified. If you can notify the recipient, you can get the message there, so you might as well. You are not inflicting it on many sites (mailing lists excepted, which my earlier posting said should have stricter standards). If you are having trouble configuring news, you can use email to get help. If you are having trouble configuring email you don't have much else to fall back on, so mail systems should be somewhat tolerant and still try to deliver. On the other hand, recipients who receive bad mail, and cannot reply directly, should complain so that the sender is pressured to fix his broken mail system. >So do I. All that I ask is that they report errors. They do. You just don't like whom they report the errors to. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
kucharsk@solbourne.com (William Kucharski) (06/08/91)
In article <oT8633w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >So tell me: why is it OK to silently throw away news a person has written, >but not OK to silently throw away mail a person has written? Because mail is generally sent to one person or a specific group of people, not to thousands of sites all over the world. If mail bounces back to you from one site or a group of sites you specifically sent to, that's fine. But I'd rather not have every C News sites send me a mail message. -- | William Kucharski, Solbourne Computer, Inc. | Opinions expressed above | Internet: kucharsk@Solbourne.COM | are MINE alone, not those | uucp: ...!{boulder,sun,uunet}!stan!kucharsk | of Solbourne... | Snail Mail: 1900 Pike Road, Longmont, CO 80501 | "It's Night 9 With D2 Dave!"
karish@mindcraft.com (Chuck Karish) (06/08/91)
In article <oT8633w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >rickert@mp.cs.niu.edu (Neil Rickert) writes: >> Mail is very different from news. > >So tell me: why is it OK to silently throw away news a person has written, >but not OK to silently throw away mail a person has written? I think it's reasonable to set a higher standard (shift more responsibility to the originator) in return for the privilege of broadcast propagation of one's words. >> I strongly support Henry and Geoff in maintaining high standards for news. > >So do I. All that I ask is that they report errors. Not long ago you were demanding that they fix errors. I'm glad this discussion is a bit more focused now. I haven't heard anyone take a principled stand against reporting errors. When a mechanism is found that is robust and inexpensive in terms of net and administrative resources I'm sure it will be implemented. -- Chuck Karish karish@mindcraft.com Mindcraft, Inc. (415) 323-9000
aipdc@castle.ed.ac.uk (Paul Crowley) (06/08/91)
In article <1991Jun6.160446.1310@ecl.psu.edu> fenner@jazz.psu.edu (Bill Fenner) writes: >no ambiguity about whether bouncing it will create huge amounts of mail >to the sender; This particular strawman is very, very dead already, you really didn't need to help garotte him. ____ \/ o\ Paul Crowley aipdc@castle.ed.ac.uk \ / /\__/ Part straight. Part gay. All queer. \/ "I say we kill him and eat his brain." "That's not the solution to _every_ problem, you know!" -- Rudy Rucker
mrm@optigfx.optigfx.com (Mike Murphy) (06/08/91)
In article <oT8633w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >rickert@mp.cs.niu.edu (Neil Rickert) writes: >> In article <m6P3324w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * >> >If a mail message is not correctly addressed, there are damn good reasons fo >> >*not* forwarding it, as has already been explained in excruciating detail. >> >Therefore we should silently delete badly-addressed mail messages. Eventuall >> >people sending mail will correct their software and learn to be more careful >> >so that erroneous messages do not occur. >> >> Mail is very different from news. > >So tell me: why is it OK to silently throw away news a person has written, >but not OK to silently throw away mail a person has written? Because news is by convention not critical. Because news is flood saturation transfer. It is, for example, perfectly OK for a site to have a manager that says, "Argggh, this ferbensplotz 7 has trashed its inode table again, clean out the news spool area, and try again." And then remove all the articles, rebuild history, active, and start over. Without telling anyone. Without notifying the authors that their articles have been trashed. Without passing the articles on to the sites that the ferbensplotz 7 feeds so dutifully. It's OK to do that. It's to be expected. It's why you have multiple feeds. It's why you feed multiple sites. Again, as an example, here at work, where I toil endlessly, we feed locally generated articles to uunet. Direct. So they'll get wide propagation. Quickly. Or sort of quickly. Now you know it is OK for a site to drop your article silently. It really is. OK for It is also OK for a site to not carry a particular distribution or newsgroup. This means that the net may be unconnected with respect to an article that you may post. Portions of the net may not get your article. This is illustrated by a simple drawing yoursite-{cloud_1 of hosts that is connected with respect to a group X} | onehost not connected with respect to X, i.e., they drop group X and don't propagate it. | {cloud_2 of hosts that is connected with respect to a group X} Cloud_2 doesn't get your postings to group X. No one needs to or is required to or is bound to tell you about it. You can get around the problem by cross posting to a group that is carried nearly everywhere, e.g., news.admin. This is considered bad form :-) You would be flamed for so doing. This is not the case for mail. I disagree with Neil. A bounce is a politeness. Sometimes the bounces don't make it. If I get a bounce notification or a bounce of a bounce back, I'll try really hard to make sure that the sender gets a report that mail isn't making it through. So they can try another route. Try hard as in human intervention, followed by telephone call if required. The mail is only going the one path, I'm to assume. I don't need to convince Neil that my view is right and his is wrong. BTW, mine isn't right and his wrong. Mine may be more polite. He has less time per mail message bounced at his site. If he had more time, who knows, he still might be as crusty as he is, maybe not :-) If the bounce is a bounce to nobody@nobody.com because somebody didn't bother to edit some canned header, I don't try very hard to let 'em know about a bounce. Again, News is different. > >> I strongly support Henry and Geoff in maintaining high standards for news. > >So do I. All that I ask is that they report errors. > They do. It (C News) does. Possibly to someone who could care less, a site administrator at a site far from you :-( I don't let you know when a batch of incoming news is dropped for lack of spool space that may contain an article that was written by you. Do you want me to start doing so? :-) I'd extend the offer to everyone on the net, but I've limited resources, as do the systems that pass mail for us, sorry... > >mathew > > -- Mike Murphy mrm@Optigfx.COM ucsd!optigfx!mrm +1 619 292 6060 x 265 Optigraphics Corporation 9339 Carroll Park Drive San Diego, CA 92121 The opinion(s) expressed above are mine and not those of my employer.
csu@alembic.acs.com (Dave Mack) (06/08/91)
In article <m6P3324w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >csu@alembic.acs.com (Dave Mack) writes: >> Here's a clue for you, mathew: we are not obligated to ensure that >> *any* article gets propagated. > >Well, no. But if you feed any other site, that site is probably under the >impression that it is getting a feed of all the articles in the newsgroups it >requests. I do hope you will explain that you are, in fact, deliberately and >unnecessarily silently discarding articles whether they like it or not. If the sites I feed are under this impression, they are sadly mistaken. I have a 60 MB news partition on this machine. If it fills up and news gets dropped on the floor, that's life. If I run out of inodes, I expire everything older than one day, and if some of it hasn't been batched yet for transmission to my downstream sites, well, that's life. And if some bit of defective software (or defective wetware) produces a malformed header and C News chucks it into the bit bucket, that also is life. If uunet gets struck by lightning (again) and all the news that arrived in the last twelve hours vanishes, that's like life, only worse. Nor do I feel compelled to report such events. I don't have the time. "Usenet is not real life" (St. Spafford.) The occasional article lost is of no particular importance in the overall scheme of things. >Furthermore, if your attitude to people at other sites making errors is "fuck >you", then presumably you don't send bounce messages for badly-addressed >mail. Of course not. My good buddy, Mailer Daemon, does it for me. If he can. If he can't, the odds are I can't either, because the problem is either a return address that is garbage or at some other site that I have no control over. Maybe I should try to get in touch by long-distance phone call to let people know that their e-mail bounced? >> And if any article doesn't conform >> to the relevant RFCs, there are damned good reasons for *not* >> propagating it, as has already been explained to you in excruciating >> detail. > >Indeed. If you want to throw away any article which is even slightly wrong, >then that's entirely up to you. Just make sure you tell the people who think >that you are forwarding articles. Sorry, no. I have to pay real live money for my communication costs. If you can't generate articles that are acceptable to the news transport software, that is entirely and exclusively your problem. <taking off the horns and lowering the pitchfork for a moment> Mathew, I'm sure you spent a lot of time and effort on the articles you posted which were trashed by the C News site upstream of you. I'm sorry about that. But what Henry and Geoff have done with C News was the right thing to do in the long run. Maybe, eventually, we'll come up with a decent scheme to notify sites with defective software without penalizing all the intermediate sites. I hope so. -- Dave Mack net.fascist
rmk@rmkhome.UUCP (Rick Kelly) (06/10/91)
In article <Vu3y312w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >tp@mccall.com (Terry Poot) writes: >> However, as I've said, I have no problem with the transition to official >> standards, just the way it was accomplished. I do firmly agree that all >> software that generates non-compliant headers should be fixed. I disagree >> that a reasonable and effective way to accomplish that is to simply ignore >> all news from sites that run such software, without bothering to tell them >> about it. > >This is my position as well. > >> On the other hand, I've said that I don't know how many times, and the >> responses are generally on another tangent, so I give up. > >I have no plans to give up, because (a) I believe that C News's current >behaviour is damaging to the people who use Usenet, and (b) I'm exceptionally >stubborn when fighting things which I genuinely believe to be wrong. > >By the latter I mean that those who want me to shut up about this issue will >have to convince me that C News's behaviour is correct; I won't just give up. I have listened to this debate for a while, and I am curious about one aspect. Who pays for email into your site? I understand it is fairly expensive to send email to the UK. If 100 postings from your site each produce 500 bad header email messages in one day, who pays for this? I believe Henry and Geoff had monetary concerns in mind when they did the most recent version of Cnews. Rick Kelly rmk@rmkhome.UUCP frog!rmkhome!rmk rmk@frog.UUCP
koerber.sin@sni.de (Mathias Koerber) (06/10/91)
In article <1991Jun7.160730.11982@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes: >In article <oT8633w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >>rickert@mp.cs.niu.edu (Neil Rickert) writes: >>> Mail is very different from news. >> >>So tell me: why is it OK to silently throw away news a person has written, >>but not OK to silently throw away mail a person has written? > > Firstly C-news does not silently throw away news. It reports it to the >local administrator. > > Secondly email is from one person to another. One of the two should be >notified. News is from one person to many persons. One of the manny should be notified. Since it doesn't make sense to notify anyone of the meny recipients, the notification should go to: a) the sender (preferably) b) an authorized person along the way who can do something about it (some admin) But just dropping is a no-no with news as well as with mail. Mathias Koerber | S iemens | EUnet: koerber.sin@sni.de 2 Kallang Sector | N ixdorf | USA: koerber.sin@sni-usa.com S'pore 1344 | I nformation Systems | Tel: +65/7402852 | Fax: +65/7402834 Thank ME it's Friday, GOD said and created the weekend|#include <disclaimer.h>
mathew@mantis.co.uk (Giving C News a *HUG*) (06/10/91)
rickert@mp.cs.niu.edu (Neil Rickert) writes: > In article <oT8633w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *H > >So tell me: why is it OK to silently throw away news a person has written, > >but not OK to silently throw away mail a person has written? > > Firstly C-news does not silently throw away news. It reports it to the > local administrator. That is silent *to the user*. I am speaking from the perspective of a user of the system. So FROM THE PERSPECTIVE OF A USER OF THE SYSTEM, why is it OK to silently throw away news that person has written, but not OK to silently throw away mail that person has written? mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/10/91)
karish@mindcraft.com (Chuck Karish) writes: > In article <oT8633w164w@mantis.co.uk> mathew@mantis.co.uk writes: > > > I strongly support Henry and Geoff in maintaining high standards for news > > > >So do I. All that I ask is that they report errors. > > Not long ago you were demanding that they fix errors. I'm glad > this discussion is a bit more focused now. Well, actually I still think that on software engineering grounds, C News should fix errors if it can. But I've given up on that approach, because I know I'm not going to persuade the C News Worshippers to take any notice of software engineering principles. > I haven't heard anyone take a principled stand against reporting > errors. When a mechanism is found that is robust and inexpensive > in terms of net and administrative resources I'm sure it will be > implemented. Indeed. And my response is that if you can't find a robust and inexpensive way to report errors, you had better make damn sure you are able to cope with said errors. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/10/91)
csu@alembic.acs.com (Dave Mack) writes: > I have a 60 MB news partition on this machine. If it fills up and > news gets dropped on the floor, that's life. Indeed, and this is a freak event which you can't do anything about. > If I run out of inodes, > I expire everything older than one day, and if some of it hasn't > been batched yet for transmission to my downstream sites, well, that's > life. This, again, is a freak event which you can't do anything to prevent. > And if some bit of defective software (or defective wetware) > produces a malformed header and C News chucks it into the bit bucket, > that also is life. But that's something you CAN do something about. I suspect that if your news spool filled up once every ten days or so, your downstream sites *would* complain. If through some random catastrophe my postings get deleted, then I don't care. What I *do* care about is when someone *deliberately* sets up his software to *repeatedly* delete my postings in order to "punish" me for using buggy software. > >Furthermore, if your attitude to people at other sites making errors is "fuc > >you", then presumably you don't send bounce messages for badly-addressed > >mail. > > Of course not. My good buddy, Mailer Daemon, does it for me. If he can. Well, your good buddy Cnews Pedantry Daemon ought to issue error reports on a similar basis. mathew
lyda@acsu.buffalo.edu (kevin lyda) (06/10/91)
In article <1991Jun7.125118@mccall.com> tp@mccall.com (Terry Poot) writes:
->If you are trying to warn people, it makes sense to post in a forum read
->by those people, with a subject line designed to attract their attention.
->
->Subject: ATTENTION non-C news newsadmins, important warning
->
->would certainly have gotten my attention, posted to news.admin,
->news.announce.important, or news.software.anu-news. I don't read
->news.software.b because I don't run B news or C news. I suspect you don't
->read news.software.anu-news for similar reasons.
->
->Posting something about C news dropping articles with bogus headers is
->about as much warning as the legal notices in the back of the newspaper
->that few people ever read. They aren't designed to notify anybody, just
->to be able to say "we did notify the public" at a later date.
good point! i agree entirely.
in other words, when someone posts something, they should use the correct
headers so that everyone that reads the headers can understand exactly what
is going on.
DO YOU GET MY DRIFT?!
kevinigb@fulcrum.bt.co.uk (Ian G Batten) (06/11/91)
In article <Z82B433w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: > I suspect that if your news spool filled up once every ten days or so, your > downstream sites *would* complain. If through some random catastrophe my Right. And he'd get a simple answer. ``Buy me more disk or find another newsfeed.'' Yes, I do runs newsfeeds. ian
rickert@mp.cs.niu.edu (Neil Rickert) (06/11/91)
In article <F31B429w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: > >So FROM THE PERSPECTIVE OF A USER OF THE SYSTEM, why is it OK to silently >throw away news that person has written, but not OK to silently throw away >mail that person has written? The email was written and addressed to a specific recipient. The news article was put on the net on a "take it or leave it" basis. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
brian@ucsd.Edu (Brian Kantor) (06/12/91)
The patches to both B news and C news to mail all failed articles to mathew are relatively trivial, but left as an exercise for the reader. - Brian
dan@gacvx2.gac.edu (06/12/91)
Greetings, ** WARNING: Sarcasm intended: ** Why hasn't cnews been fixed to stop the rebroadcasting of news by broken sites like the fido net node currently going wild? It seems to me that filtering out/droping all or most news would fix the problem and save bandwidth too. Perhaps cnews could do something about followups as well, and don't forget crossposting. ** Seriousness: *** I would think that it would be better for the authors of cnews and other news systems to spend the time preventing their software from sending out old news, by fixing the conditions that cause old news to be sent, rather than filtering it from the net at large. -- Dan Boehlke Internet: dan@gac.edu Campus Network Manager BITNET: dan@gacvax1.bitnet Gustavus Adolphus College St. Peter, MN 56082 USA Phone: (507)933-7596
mathew@mantis.co.uk (Giving C News a *HUG*) (06/12/91)
igb@fulcrum.bt.co.uk (Ian G Batten) writes: > In article <Z82B433w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * > > I suspect that if your news spool filled up once every ten days or so, your > > downstream sites *would* complain. If through some random catastrophe my > > Right. And he'd get a simple answer. ``Buy me more disk or find another > newsfeed.'' Right. But unfortunately I can't get a non-C-news newsfeed, because at least some articles will always be lost by C News before they get to whatever site provides me with the feed. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/12/91)
rickert@mp.cs.niu.edu (Neil Rickert) writes: > In article <F31B429w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * > >So FROM THE PERSPECTIVE OF A USER OF THE SYSTEM, why is it OK to silently > >throw away news that person has written, but not OK to silently throw away > >mail that person has written? > > The email was written and addressed to a specific recipient. The news > article was put on the net on a "take it or leave it" basis. Really? Where does it say this? I don't recall anything in the news.announce.newusers postings saying "Your articles may be discarded silently before they reach the net if you make a typo". If the conditions of using Usenet really are that you have to accept 0% confidence in its reliability, then I think that ought to be spelled out. mathew
karish@mindcraft.com (Chuck Karish) (06/12/91)
In article <F31B429w164w@mantis.co.uk> mathew@mantis.co.uk writes: |rickert@mp.cs.niu.edu (Neil Rickert) writes: || In article <oT8633w164w@mantis.co.uk| mathew@mantis.co.uk (Giving C News a *H || |So tell me: why is it OK to silently throw away news a person has written, || |but not OK to silently throw away mail a person has written? || || Firstly C-news does not silently throw away news. It reports it to the || local administrator. | |That is silent *to the user*. I am speaking from the perspective of a user of |the system. | |So FROM THE PERSPECTIVE OF A USER OF THE SYSTEM, why is it OK to silently |throw away news that person has written, but not OK to silently throw away |mail that person has written? Mathew has crossed the line from annoying repetitive (though marginally productive) flaming to intellectual dishonesty. His quote of Neil Rickert's article is taken out of context, in that it ignores the second half of Neil's article which addressed this issue directly. -- Chuck Karish karish@mindcraft.com Mindcraft, Inc. (415) 323-9000
moraes@cs.toronto.edu (Mark Moraes) (06/13/91)
dan@gacvx2.gac.edu writes: >I would think that it would be better for the authors of cnews and other news >systems to spend the time preventing their software from sending out old news, >by fixing the conditions that cause old news to be sent, rather than filtering >it from the net at large. Ah, so it's now possible for news software to detect and prevent administrator error and bad gateway design. Some scenarios that are a little difficult to prevent in software: - site goes down for a while, comes up with old news batches (note: not batchfiles) still around and sends them out, since they're in a uucp queue. - someone restores a trashed partition from fairly old backups and accidentally restores some old news batches. - a gateway from a network that does not preserve message-ids goes nuts. Yes, we've seen all of these happen; some of them quite recently, in fact. Mark. -- "It's only netnews"
rickert@mp.cs.niu.edu (Neil Rickert) (06/13/91)
In article <89HF414w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >rickert@mp.cs.niu.edu (Neil Rickert) writes: >> >> The email was written and addressed to a specific recipient. The news >> article was put on the net on a "take it or leave it" basis. > >Really? Where does it say this? I don't recall anything in the >news.announce.newusers postings saying "Your articles may be discarded >silently before they reach the net if you make a typo". If your usenet neighbors were to place in their sys files (please, please) ME:all,!news.admin,!news.software.bin/all:: your articles on this subject would go nowhere. Is this supposed to be in news.announce.newusers ? However, I have a proposed solution to your problems: Attention all Cnews sites. Please place the following commands near the end of your newsdaily scripts: cd $NEWSCTL grep ' - <' log.o | \ grep -v 'duplicate$' | \ grep -v `cat whoami` | mail mathew@mantis.co.uk There. If enough admins pick this up you will soon be fully informed of all articles of yours dropped due to bad headers or dates. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
sharon@jpl-devvax.jpl.nasa.gov (Sharon Hopkins) (06/13/91)
In article <89HF414w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: > >If the conditions of using Usenet really are that you have to accept 0% >confidence in its reliability, then I think that ought to be spelled out. Spelled out where? Why? Usenet doesn't exactly have a Corporate Charter, and I don't see any warranty card enclosed in the Periodic Postings. Who should be spelling this out? Gene Spafford? He seems to be pretty busy doing his regular job these days. Henry Spencer? He sounds pretty busy, too. Chip Salzenberg? He tried to give his opinion of what Usenet does and does not offer, but his statement is currently attracting considerable discussion and disagreement, even if mostly from one person. :-) Hmm. I think I will dissent from the view that Usenet is an organization of news administrators, and present instead the Sharon Hopkins slightly slanted view of the Usenet Warranty, or lack thereof. This being Usenet, I think I shall start with an analogy (preferably a poor one :-) : When I wander into town and tack notices up on the local signboard, I don't assume that the existence of the signboard is any guarantee that I'll get a response, or even that anyone else will see what I've posted there. Maybe someone will tear my notice down, or cover it up. That isn't *nice*, but that's how things are with notice-boards. Given where I live, it's quite possible someone will grafitti over it, or alter it's appearance (somewhat in the manner of a garbled header). If my notice is defaced in some fashion, it might well get taken down by some civic-minded person with orderly habits and an eye towards public decency. My only means of checking the status of my notice is by getting a response (say, a bottle through my window, or an offer to take the cute puppy or kitten off my hands -- it sort of depends on the notice, doesn't it?), or (in the event I get *no* response) by going back to the sign-board and looking. Suppose, when I go to look at my notice, I discover that it is missing. Would it be reasonable for me to yell at the owner of the wall, who mounted a nice piece of cork-board for the purpose, for the fact that my notice disappeared? I would tend not to think so, though I'm certain at least a few people would disagree. If we want to draw the analogy even closer, then we have to also take into consideration the fact that we rarely put up the notices ourselves: some of us pay someone else to put it up for us; others of us rely on somebody's good will. Can you yell at your notice-carrier if no one responds? Or if they put it up and someone else takes it down when they aren't looking? Or if they hand it off to a third party who drops it in a garbage bin? Maybe, if they were for some reason bound to keep track of your notice for you. If they had only sort of implied that your notice would be posted, and that other people would be able to look at it, I would say you have no real right to complain if things don't go as you would have liked. To me, this scenario is not unlike posting an article on (by? in?) Usenet; I figure there's a reason why this business of setting "articles" loose all across the net is called "posting" and not "publishing". In my particular case, I hand my notices (articles) off to somebody else headed in the right general direction, who passes them along to someone else, who may or may not put them where I want. (The fact that a random bunch of people and organizations with computers are engaged in a cooperative effort to provide rapid (and often cheap) communication, does not imply to me that they will necessarily succeed (or even want to succeed) in transmitting *my* communication.) In other words, I use rn, on a system running C-News, which is administered by a couple of guys in offices down the hall. Our machine sends articles and the like to a machine a few buildings over, which is in charge of getting them off-site; once news leaves my screen, it's pretty much out of my hands. Do I get annoyed when articles I spent hours preparing never make it off our local machine? Sure I do. I get a lot more annoyed when articles I spent lots of time on get wiped out when the partition I'm in overflows and vi can't write to disk, but that happens often enough that I've learned to cope. :-) But when I go long enough without getting a reply to some article I've posted, my first response is not generally to flood the net with post after post (after post) complaining, demanding apologies, and the like; for one thing, it seems like an insufficiently productive way to spend my time, though that may just be because I'm too shy to yell long enough to get anyone to do what I want (read up on the parable of the unjust judge :-). Instead, I usually just wander into the SA's office and tell him my news doesn't seem to be getting out. Eventually someone usually figures out what went wrong and sees about getting it fixed. Sometimes they even thank me for noticing a problem. Sometimes, they say (more or less) "tough luck". Meanwhile, I generally manage to survive. You see, I don't think having an account on this particular machine gives me some kind of inherent right to post articles: if the SA (or management) pulls the plug, that's it. When the disk crashes, or the next machine in line stops working, some articles get lost. Oh well. If my articles get disappeared somewhere down the line, seems to me that's more or less in the same category with things like disk crashes: I can't fix it, and somebody else may or may not fix it for me. If I get really desperate to express myself, I can always write a letter to the editor, or call someone who cares, or wander into town to tack up a sign. Usenet is mostly just a funner, faster way to carry on a conversation; as with other forms of conversation, there are no guarantees that anyone else will be able to hear (or understand) you. There's even less guarantee anyone will listen. Usenet Product Label: "This service carries no warranties, express or implied: try not to take it too personally." Sharon Hopkins sharon@jpl-devvax.jpl.nasa.gov
dylan@ibmpcug.co.uk (Matthew Farwell) (06/13/91)
In article <1991Jun12.180726.4461@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes: > Attention all Cnews sites. Please place the following commands near the >end of your newsdaily scripts: > > cd $NEWSCTL > grep ' - <' log.o | \ > grep -v 'duplicate$' | \ > grep -v `cat whoami` | mail mathew@mantis.co.uk > > There. If enough admins pick this up you will soon be fully informed of all >articles of yours dropped due to bad headers or dates. Two things. One, this won't actually work, because we (mantis's feed) run the latest Cnews, so any bad articles shouldn't get out onto the rest of the net. Two, more importantly, encouraging this kind of net.terrorism isn't very nice, IMHO. We are on quite an expensive link to ukc, so not only will this mean that we have to suffer, but quite a bit of uknet has to suffer in a degradation of performance from ukc. Before you say this was a joke, I'd like to say that advocating this sort of thing is not really a good idea, joke or not. Dylan. -- Matthew J Farwell: dylan@ibmpcug.co.uk || ...!uunet!ukc!ibmpcug!dylan But you're wrong Steve. You see, its only solitaire.
aipdc@castle.ed.ac.uk (Paul Crowley) (06/13/91)
In article <676736110.29234@mindcraft.com> karish@mindcraft.com (Chuck Karish) writes: >Mathew has crossed the line from annoying repetitive (though >marginally productive) flaming to intellectual dishonesty. You know, I so wish we could stop being repetitive. We have to keep saying things like "No, we don't think that sites should mail reports to the posters of bad articles" and "Who is to blame for the bad article has nothing to do with the case" over and over again. Perhaps if people could stop dredging up the same old strawmen we could stop refuting them and move on to productive debate. As for intellectual dishonesty, Neil's first point is simply incorrect, and remains incorrect within the context of the article. Mathew pointed the error out. If he'd done a point-by-point refutation of the article, he would have been accused of repetition. Accusations of repetition play an interesting part in debate. Polititians use them to avoid questions. "So, what do you intend to do about it?" "I'm glad you asked me that question. The situation today is simply unnacceptable, and it was one of my election pledges to resolve it. I intend to keep that pledge." "So, what do you intend to do about it?" "We're setting up a subcommittee to take action on it. We polled local residents and over seventy percent said that action is needed, and it is a scandal that the opposition party took no action at all." "So, what do you intend to do about it?" "Please, stop being so repetitive." Sad to say, repetition can be very necessary in some debates. As Prince said "There's joy in repetition. There's joy in repetition." ____ \/ o\ Paul Crowley aipdc@castle.ed.ac.uk \ / /\__/ Part straight. Part gay. All queer. \/ "I say we kill him and eat his brain." "That's not the solution to _every_ problem, you know!" -- Rudy Rucker
aipdc@castle.ed.ac.uk (Paul Crowley) (06/13/91)
In article <1991Jun12.180726.4461@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes: > Attention all Cnews sites. Please place the following commands near the >end of your newsdaily scripts: After all, why argue when you can threaten? Oh, I see, you're just making a point that you can't mail errors to the originators of articles. We never realised that, did we Mathew? Oh, I see, it's just a joke. Neil, if you continue to disagree with me, I'll come over and cut your balls off with a chainsaw. Ha ha, just a joke. Funny me. ____ \/ o\ Paul Crowley aipdc@castle.ed.ac.uk \ / /\__/ Part straight. Part gay. All queer. \/ "I say we kill him and eat his brain." "That's not the solution to _every_ problem, you know!" -- Rudy Rucker
rickert@mp.cs.niu.edu (Neil Rickert) (06/13/91)
In article <1991Jun13.073739.1131@ibmpcug.co.uk> dylan@ibmpcug.CO.UK (Matthew Farwell) writes: >In article <1991Jun12.180726.4461@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes: >> cd $NEWSCTL >> grep ' - <' log.o | \ >> grep -v 'duplicate$' | \ >> grep -v `cat whoami` | mail mathew@mantis.co.uk > >Two, more importantly, encouraging this kind of net.terrorism isn't very >nice, IMHO. We are on quite an expensive link to ukc, so not only will >this mean that we have to suffer, but quite a bit of uknet has to suffer >in a degradation of performance from ukc. > >Before you say this was a joke, I'd like to say that advocating this >sort of thing is not really a good idea, joke or not. I will admit that not all news readers recognize jokes when they see them. That, after all, is how this particular subject line started several centuries ago. But, consider: My "proposal", since it called for modification of news software, could only be implemented by news administrators at sites running Cnews with the latest patches. And even then, it would require some minimal level of technical skills in writing and modifying shell scripts to implement. One should be able to assume that those capable of implementing the "proposal" have a slightly higher intelligence and a slightly higher sense of responsibility than the average netnews reader, and although I have no doubt that there are some exceptions, it is highly unlikely that there are enough to cause significant net traffic. When I wrote that, I did hope it would make Mathew (no, not Matthew) think more seriously about the real implications of what he has been continuously demanding for an interminable length of time. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
mathew@mantis.co.uk (Giving C News a *HUG*) (06/13/91)
karish@mindcraft.com (Chuck Karish) writes: > In article <F31B429w164w@mantis.co.uk> mathew@mantis.co.uk writes: > |So FROM THE PERSPECTIVE OF A USER OF THE SYSTEM, why is it OK to silently > |throw away news that person has written, but not OK to silently throw away > |mail that person has written? > > Mathew has crossed the line from annoying repetitive (though > marginally productive) flaming to intellectual dishonesty. His > quote of Neil Rickert's article is taken out of context, in > that it ignores the second half of Neil's article which > addressed this issue directly. It wasn't "intellectual dishonesty". I didn't look at the second half of his article, since it was apparent from the first half that he had misunderstood that I was referring to the perspective of a user of the system. Since you insist, here's a reply to the entire article: In <1991Jun7.160730.11982@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes: > Firstly C-news does not silently throw away news. It reports it to the >local administrator. From the perspective of a user of the system, that does not constitute reporting the error. > Secondly email is from one person to another. One of the two should be >notified. News is from one person to several. Either the sender or the recipients should be notified. > If you can notify the recipient, you can get the message there, >so you might as well. If you can deliver a news message, then you might as well, even if it contains errors. > You are not inflicting it on many sites (mailing >lists excepted, which my earlier posting said should have stricter >standards). News is inflicted on many sites as it is. I don't think inflicting one error report instead of each bad article is a retrograde step. > If you are having trouble configuring news, you can use email to get help. >If you are having trouble configuring email you don't have much else to >fall back on, so mail systems should be somewhat tolerant and still try >to deliver. Rubbish. You can fall back to telephone or postal mail. The fact that there is an alternative to a system is not a good excuse for making that system unreliable. > On the other hand, recipients who receive bad mail, and cannot >reply directly, should complain so that the sender is pressured to fix his >broken mail system. Recipients who receive bad news and cannot handle it should complain so that the sender is pressured to fix his broken news system. >> All that I ask is that they report errors. > > They do. You just don't like whom they report the errors to. From the viewpoint of a user of the system, they do not report errors. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/13/91)
rickert@mp.cs.niu.edu (Neil Rickert) writes: > In article <89HF414w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * > >rickert@mp.cs.niu.edu (Neil Rickert) writes: > >> The email was written and addressed to a specific recipient. The news > >> article was put on the net on a "take it or leave it" basis. > > > >Really? Where does it say this? I don't recall anything in the > >news.announce.newusers postings saying "Your articles may be discarded > >silently before they reach the net if you make a typo". > > If your usenet neighbors were to place in their sys files (please, please) > > ME:all,!news.admin,!news.software.bin/all:: > > your articles on this subject would go nowhere. Is this supposed to be in > news.announce.newusers ? Whether a site receives a newsgroup is visible to readers at that site. Whether the site silently discards postings containing typos is not visible to people at that site. The two situations are therefore entirely different. > However, I have a proposed solution to your problems: > > Attention all Cnews sites. Please place the following commands near the > end of your newsdaily scripts: [ Dull UNIX script deleted ] Wow, what a screamingly amusing joke. It was almost as amusing that time as it was the last time someone made it during this discussion, or the time before, or the time before that. Might I suggest that your time would be better spent writing code to fix C News rather than writing geeky UNIX-hacker jokes? mathew
chip@tct.com (Chip Salzenberg) (06/14/91)
According to mathew@mantis.co.uk (Giving C News a *HUG*):
>I am speaking from the perspective of a user of the system.
New users don't understand the issues involved in error notification.
If you design everything from the perspective of a naive user, you get
a naive system.
--
Chip Salzenberg at Teltronics/TCT <chip@tct.com>, <uunet!pdn!tct!chip>
"You can call Usenet a democracy if you want to. You can call it a
totalitarian dictatorship run by space aliens and the ghost of Elvis.
It doesn't matter either way." -- Dave Mackchip@tct.com (Chip Salzenberg) (06/14/91)
According to mathew@mantis.co.uk (Giving C News a *HUG*): >Well, actually I still think that on software engineering grounds, C News >should fix errors if it can. I heartily disagree. The venerable software engineering principle that urges to "be conservative in what you generate" makes obvious the wrongness of attempting to read a human's mind and pass the results off as valid. Well-engineered software is simple and reliable. And as the Jargon File points out, "DWIM (Do What I Mean) is incredibly hairy." >If you can't find a robust and inexpensive way to report errors, you had >better make damn sure you are able to cope with said errors. Of course, "cope with" sometimes translates to "drop into /dev/null". -- Chip Salzenberg at Teltronics/TCT <chip@tct.com>, <uunet!pdn!tct!chip> "You can call Usenet a democracy if you want to. You can call it a totalitarian dictatorship run by space aliens and the ghost of Elvis. It doesn't matter either way." -- Dave Mack
chip@tct.com (Chip Salzenberg) (06/14/91)
According to mathew@mantis.co.uk (Giving C News a *HUG*): >What I *do* care about is when someone *deliberately* sets up his software >to *repeatedly* delete my postings in order to "punish" me for using buggy >software. "Punish"? Surely you're getting rather paranoid here, Mathew. Tossing invalid articles is simply neighborly. A badly-formed article is a failure waiting to happen. That my neighbors drop them on the floor bothers me not one whit. Don't get me wrong. Failure notices are a Good Thing *when practical*. But even if they turn out not to be practical, I will still support the lossage of non-RFC articles. -- Chip Salzenberg at Teltronics/TCT <chip@tct.com>, <uunet!pdn!tct!chip> "You can call Usenet a democracy if you want to. You can call it a totalitarian dictatorship run by space aliens and the ghost of Elvis. It doesn't matter either way." -- Dave Mack
henry@zoo.toronto.edu (Henry Spencer) (06/14/91)
In article <Vu3y312w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >I have no plans to give up... You might as well stop orating about it, however, because you're talking to yourself. Most everybody -- including Geoff and I -- has stopped reading your messages. We've heard everything you've got to say. -- "We're thinking about upgrading from | Henry Spencer @ U of Toronto Zoology SunOS 4.1.1 to SunOS 3.5." | henry@zoo.toronto.edu utzoo!henry
csu@alembic.acs.com (Dave Mack) (06/14/91)
In article <Z82B433w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >csu@alembic.acs.com (Dave Mack) writes: >> I have a 60 MB news partition on this machine. If it fills up and >> news gets dropped on the floor, that's life. > >Indeed, and this is a freak event which you can't do anything about. > >> If I run out of inodes, >> I expire everything older than one day, and if some of it hasn't >> been batched yet for transmission to my downstream sites, well, that's >> life. > >This, again, is a freak event which you can't do anything to prevent. One of the most common causes of these "freak" events is when some idiot screws around with news headers and then dumps a huge volume of ancient articles into the net. This is one of things that Cnews now prevents. >> And if some bit of defective software (or defective wetware) >> produces a malformed header and C News chucks it into the bit bucket, >> that also is life. > >But that's something you CAN do something about. I could, yes. But why should I? >I suspect that if your news spool filled up once every ten days or so, your >downstream sites *would* complain. If through some random catastrophe my >postings get deleted, then I don't care. What I *do* care about is when >someone *deliberately* sets up his software to *repeatedly* delete my postings >in order to "punish" me for using buggy software. No one is "punishing" you, unless whoever produced your defective posting software knew it was defective. Malformed articles can cause problems for the entire net. The Cnews patches help reduce those problems. This is A Good Thing. It helps solve the problem. Your buggy software was part of the cause of the problem. So, yes, you suffered. Most unfortunate. But not a reason for the rest of the net to revert to having to deal with spasms of ancient news articles, nor for us to have to deal with the possibility of floods e-mail notices to sites generating bad headers. >> >Furthermore, if your attitude to people at other sites making errors is "fuc >> >you", then presumably you don't send bounce messages for badly-addressed >> >mail. >> >> Of course not. My good buddy, Mailer Daemon, does it for me. If he can. > >Well, your good buddy Cnews Pedantry Daemon ought to issue error reports on a >similar basis. There is a certain amount of effort going on right now to find a reasonable way to do this. As yet, no one has found a way that is both guaranteed to return an error report and also guaranteed not to flood the net with them. When someone does find such a method, it will probably be incorporated into C News. Until then, people with buggy posting software will have to rely on the kindness of strangers. -- Dave Mack
rickert@mp.cs.niu.edu (Neil Rickert) (06/14/91)
In article <sTHH412w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >In <1991Jun7.160730.11982@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) >writes: >> If you are having trouble configuring news, you can use email to get help. >>If you are having trouble configuring email you don't have much else to >>fall back on, so mail systems should be somewhat tolerant and still try >>to deliver. > >Rubbish. You can fall back to telephone or postal mail. Have you ever tried sending a 500 line context diff through the telephone? Or redirecting the audio output of your telephone answerer into patch? -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
herrickd@iccgcc.decnet.ab.com (06/14/91)
In article <wu2N39w164w@mantis.co.uk>, mathew@mantis.co.uk (CNEWS MUST DIE!) writes: > > Right. But the users who happen to be using the software written by the > person who didn't follow the RFC are innocent bystanders; and innocent > bystanders should not be punished. > But, Mathew, that is the way the world works. When the control systems, computerised and human, fail and two trains in the underground run into each other, it is not the author of the software or the operator in the control center who goes to the hospital. It is always the innocent bystanders who get punished. (Though the news this am suggests some of the non-innocent may get their turn in the matter of flight 103.) The people who bought and paid for Ford Escorts with poorly placed gas tanks (how many of your innocent bystanders bought and paid for their news posting software) were "innocent bystanders" the way you are using the word. Some of them had a very hot time of it. Maybe a more useful function here would be to identify the bad software. What posting software were you using, Mathew, when you were dumping non-compliant articles onto the poor unsuspecting network? How about it, folks? Program name and version number and date. What software posts bad articles. If that data becomes known to the people who are using it, they can switch to a version that works or to a different program entirely. dan herrick herrickd@iccgcc.decnet.ab.com
mathew@mantis.co.uk (Giving C News a *HUG*) (06/14/91)
Wow. Kevin Darcy's flaming me. I *must* be right. kevin@rotag.mi.org (Kevin Darcy) writes: > Mathew: it is not the job of C News to de-louse every bad header it sees, nor > to yak-yak to humans about the sad state of posting software. It's purpose is > to transfer news to and from disks and networks as smoothly and efficiently a > possible. Right. And for six months, it was failing to transfer news articles I posted. Smoothly and efficiently don't even enter into it, it was simply failing in its purpose. > C NEWS IS A NEWS *TRANSPORT*, NOT A NEWS *SCHOOLMARM*!! Right. So why does it complain about errors that are unimportant? So long as it has enough information to deliver the article, why the hell does it schoolmarm us with nit-picking about the header syntax? > Given the obvious advantages and increasing popularity of C News over other > news transports, and the fact that it follows the RFC in this regard, I > consider it perfectly reasonable to expect sites with broken posting software > to adapt to C News, rather than the other way around. So do I. However, I happen to take the rather sensible view that sites with broken software are unlikely to spontaneously mutate into C News compliant sites. Some human intervention at the appropriate site will be required, and hence someone at that site needs to be reliably informed about errors detected by C News. Hence C News has to have an error-reporting mechanism. I think you ought to re-think whose side you're arguing for. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/14/91)
rickert@mp.cs.niu.edu (Neil Rickert) writes: [ Unfunny "joke" about mailing error messages deleted ] > > When I wrote that, I did hope it would make Mathew (no, not Matthew) think > more seriously about the real implications of what he has been continuously > demanding for an interminable length of time. I presume that you suffer from some sort of reading comprehension problem, or perhaps that you have some bizarre obsession with straw-man arguments. You should know by now that I am not proposing that mail be used as the primary means of reporting errors. Nor have I even suggested it as a possibility for several weeks. The only mail-based news error reporting I currently believe to be a sensible idea is that which would be performed by one or two backbone sites. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/14/91)
rmk@rmkhome.UUCP (Rick Kelly) writes: > In article <Vu3y312w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * > >By the latter I mean that those who want me to shut up about this issue will > >have to convince me that C News's behaviour is correct; I won't just give up > > I have listened to this debate for a while, Judging from your question, I am supremely doubtful as to the accuracy of that statement. > If 100 postings from your site each > produce 500 bad header email messages in one day, who pays for this? As far as I am aware, nobody is suggesting that every C News site send email to the source of a bad article. Quite what relevance your question has to the debate at hand is therefore unclear. mathew
bengtl@maths.lth.se (Bengt Larsson) (06/15/91)
In article <1991Jun14.084539.4864@iccgcc.decnet.ab.com> herrickd@iccgcc.decnet.ab.com writes: >Maybe a more useful function here would be to identify the bad software. >What posting software were you using, Mathew, when you were dumping >non-compliant articles onto the poor unsuspecting network? Yes. How about it, Mathew? Which software were you using which got you into problems? Name and version of the software? You were using an Atari ST, right? >If that data becomes known to the people >who are using it, they can switch to a version that works or to a >different program entirely. Exactly. -- Bengt Larsson - Dep. of Math. Statistics, Lund University, Sweden Internet: bengtl@maths.lth.se SUNET: TYCHE::BENGT_L
chip@tct.com (Chip Salzenberg) (06/17/91)
According to mathew@mantis.co.uk (Giving C News a *HUG*): >For six months, [C News] was failing to transfer news articles I posted. >Smoothly and efficiently don't even enter into it, it was simply failing in >its purpose. Its purpose is to transmit *correct* articles. It succeeds. -- Chip Salzenberg at Teltronics/TCT <chip@tct.com>, <uunet!pdn!tct!chip> "You can call Usenet a democracy if you want to. You can call it a totalitarian dictatorship run by space aliens and the ghost of Elvis. It doesn't matter either way." -- Dave Mack
mathew@mantis.co.uk (Giving C News a *HUG*) (06/17/91)
henry@zoo.toronto.edu (Henry Spencer) replies to one of my articles, saying: > Most everybody -- including Geoff and I -- has stopped reading you > messages. Do you make a habit of replying to articles you haven't read? mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/17/91)
csu@alembic.acs.com (Dave Mack) writes: > Malformed articles can cause > problems for the entire net. The Cnews patches help reduce those > problems. This is A Good Thing. It helps solve the problem. The Problem is that there is buggy software. The C News silent-article-dropping does nothing to solve the problem, it merely hides the problem from the rest of the net. I'm suggesting error-reporting because that would actually treat the cause of the problem, rather than the symptoms. > So, yes, you suffered. > Most unfortunate. But not a reason for the rest of the net to revert > to having to deal with spasms of ancient news articles, nor for us > to have to deal with the possibility of floods e-mail notices to sites > generating bad headers. NOBODY IS SUGGESTING FLOODS OF E-MAIL. NOBODY IS SUGGESTING ALLOWING OLD NEWS TO PROPOGATE. I don't know, the same old strawman arguments YET AGAIN. Are you trying to bore me into submission? > There is a certain amount of effort going on right now to find a > reasonable way to do this. As yet, no one has found a way that is both > guaranteed to return an error report and also guaranteed not to flood > the net with them. Firstly, is any effort being expended on fixing the problem? Henry seems to have gone silent on the entire issue. Secondly, no method is "guaranteed" to return an error report. It's a matter of implementing a method which is reasonably likely to succeed. I believe that news-based error reports with a modified message-ID are an adequate and safe solution; they are likely to get to the site or gateway sending out the duff articles, and they will not flood the net. If anyone has any objections to such a system, please would he voice them, as I've not seen any proper objections yet. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/17/91)
herrickd@iccgcc.decnet.ab.com writes: > In article <wu2N39w164w@mantis.co.uk>, mathew@mantis.co.uk (CNEWS MUST DIE!) > > Right. But the users who happen to be using the software written by the > > person who didn't follow the RFC are innocent bystanders; and innocent > > bystanders should not be punished. > > But, Mathew, that is the way the world works. [...] > It is always the innocent bystanders who get punished. So?! That doesn't make it *right* to punish innocent bystanders. And since it is in our power *not* to punish them so severely, surely we *should* aim to be a little more friendly towards them? mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/17/91)
bengtl@maths.lth.se (Bengt Larsson) writes: > In article <1991Jun14.084539.4864@iccgcc.decnet.ab.com> > herrickd@iccgcc.decnet.ab.com writes: > >Maybe a more useful function here would be to identify the bad software. > >What posting software were you using, Mathew, when you were dumping > >non-compliant articles onto the poor unsuspecting network? > > Yes. How about it, Mathew? Which software were you using which got you > into problems? Name and version of the software? You were using an > Atari ST, right? I've already reported the problem to comp.sys.atari.st. In fact, that was the first thing I did when I found out about it. The author has also been informed. mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/17/91)
jmaynard@thesis1.med.uth.tmc.edu (Jay Maynard) writes: > In article <Vu3y312w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * > >By the latter I mean that those who want me to shut up about this issue will > >have to convince me that C News's behaviour is correct; I won't just give up > > I have no reason to believe that you'll ever be convinced of anything of > the sort. Do you have any reason to believe that I won't? If not, aren't you just condemning me for no good reason? mathew
mathew@mantis.co.uk (Giving C News a *HUG*) (06/17/91)
csu@alembic.acs.com (Dave Mack) writes: > One of the most common causes of these "freak" events is when some > idiot screws around with news headers and then dumps a huge volume > of ancient articles into the net. This is one of things that Cnews > now prevents. Nobody is suggesting that it should not continue to do so. > >> And if some bit of defective software (or defective wetware) > >> produces a malformed header and C News chucks it into the bit bucket, > >> that also is life. > > > >But that's something you CAN do something about. > > I could, yes. But why should I? For the same reason you presumably issue bounce messages for failed mail. mathew
billd@fps.com (Bill Davidson) (06/18/91)
I keep chanting to myself: "I will stay out of this stupid flame war. I will ignore Mathew's ranting" "I will stay out of this stupid flame war. I will ignore Mathew's ranting" "I will stay out of this stupid flame war. I will ignore Mathew's ranting" "I will stay out of this stupid flame war. I will ignore Mathew's ranting" "I will stay out of this stupid flame war. I will ignore Mathew's ranting" "I will stay out of this stupid flame war. I will ignore Mathew's ranting" Arrrrrrrgghhh! I can't help help myself. Why do I care? I don't know. In article <o4FJ44w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >kevin@rotag.mi.org (Kevin Darcy) writes: >> Mathew: it is not the job of C News to de-louse every bad header it sees, nor >> to yak-yak to humans about the sad state of posting software. It's purpose is >> to transfer news to and from disks and networks as smoothly and efficiently a >> possible. > >Right. And for six months, it was failing to transfer news articles I posted. >Smoothly and efficiently don't even enter into it, it was simply failing in >its purpose. SIX MONTHS!?!?! The last part of the patch that started dropping articles was issued on March 25 (less than 3 months ago). This flame war has been going for a couple of months now. Assuming that your feed sites patched ASAP, I think you figured it out in less than a month. Please get a sense of proportion. >> C NEWS IS A NEWS *TRANSPORT*, NOT A NEWS *SCHOOLMARM*!! > >Right. So why does it complain about errors that are unimportant? So long as >it has enough information to deliver the article, why the hell does it >schoolmarm us with nit-picking about the header syntax? How many times do we have to tell you? It's main complaint is about important errors which DO AFFECT DELIVERY wether you understand it or not. Yes "Date:" does affect delivery. In fact, transport problems related to "Date:" and people bitching about them were one of the main reasons for the new behaviour. You obviously haven't been listening. Dropping articles with bad headers has solves some major transport problems. >> Given the obvious advantages and increasing popularity of C News over other >> news transports, and the fact that it follows the RFC in this regard, I >> consider it perfectly reasonable to expect sites with broken posting software >> to adapt to C News, rather than the other way around. > >So do I. However, I happen to take the rather sensible view that sites with >broken software are unlikely to spontaneously mutate into C News compliant >sites. Some human intervention at the appropriate site will be required, >and hence someone at that site needs to be reliably informed about errors >detected by C News. Nope. If your articles get dropped that's your problem and nobody else's. News is a dispersion transport with NO GUARANTEES. When I first set up news, I posted articles to misc.test to make sure that articles were getting out to major sites. I still do it on occasion, especially when I update my news software or I know a popular patch has gone out to a lot of major sites. I am responsible for my own software and nobody else. >Hence C News has to have an error-reporting mechanism. To the local sysadmin yes. To remote admins or users no. In fact, I consider it highly undesirable. News is not mail. >I think you ought to re-think whose side you're arguing for. I've thought about it a lot. I think that the decisions made for Cnews were the right ones. --Bill Davidson
mathew@mantis.co.uk (Giving C News a *HUG*) (06/18/91)
chip@tct.com (Chip Salzenberg) writes: > According to mathew@mantis.co.uk (Giving C News a *HUG*): > >For six months, [C News] was failing to transfer news articles I posted. > >Smoothly and efficiently don't even enter into it, it was simply failing in > >its purpose. > > Its purpose is to transmit *correct* articles. It succeeds. No, the purpose of the news transport software is to correctly transmit the news articles which it receives. C News failed in that purpose. mathew
henry@zoo.toronto.edu (Henry Spencer) (06/18/91)
In article <u4wo42w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >> Most everybody -- including Geoff and I -- has stopped reading you >> messages. > >Do you make a habit of replying to articles you haven't read? It's uncommon, but I do it when there is hope of thereby reducing the noise level. :-) I haven't actually got you in my kill file yet -- I tend to use kill files rather selectively -- but a posting from you usually gets no more than one brief glance. -- "We're thinking about upgrading from | Henry Spencer @ U of Toronto Zoology SunOS 4.1.1 to SunOS 3.5." | henry@zoo.toronto.edu utzoo!henry
dylan@ibmpcug.co.uk (Matthew Farwell) (06/18/91)
In article <18522@celit.fps.com> billd@fps.com (Bill Davidson) writes: >>Right. And for six months, it was failing to transfer news articles I posted. >>Smoothly and efficiently don't even enter into it, it was simply failing in >>its purpose. >SIX MONTHS!?!?! The last part of the patch that started dropping >articles was issued on March 25 (less than 3 months ago). This flame >war has been going for a couple of months now. Assuming that your feed >sites patched ASAP, I think you figured it out in less than a month. >Please get a sense of proportion. We didn't. I waited about a month before upgrading, to make sure that the patches were going to be stable. Dylan. -- Matthew J Farwell: dylan@ibmpcug.co.uk || ...!uunet!ukc!ibmpcug!dylan But you're wrong Steve. You see, its only solitaire.
herrickd@iccgcc.decnet.ab.com (06/19/91)
In article <F31B429w164w@mantis.co.uk>, mathew@mantis.co.uk (Giving C News a *HUG*) writes: > > So FROM THE PERSPECTIVE OF A USER OF THE SYSTEM, why is it OK to silently > throw away news that person has written, but not OK to silently throw away > mail that person has written? > Hey, Henry! Here's a new spec. Let's change cnews so that every time it drops an article on the floor it tries to find a mail message with the same author and drop it, too. dan
chip@tct.com (Chip Salzenberg) (06/19/91)
According to mathew@mantis.co.uk (Giving C News a *HUG*): >The Problem is that there is buggy software. The C News >silent-article-dropping does nothing to solve the problem, it merely hides >the problem from the rest of the net. The word is not "hide", it's "protect." As in, "C News protects me and my neighbors from bad articles that arrive at my site." >I'm suggesting error-reporting because that would actually treat the cause of >the problem, rather than the symptoms. Hardly. It enlarges the set of people who are notified about the problem. But E-Mail notices can be ignored just as easily as log entries. >I believe that news-based error reports with a modified message-ID are an >adequate and safe solution; they are likely to get to the site or gateway >sending out the duff articles, and they will not flood the net. Kent's concern about a local flood of bad news generating a global flood of error messages may need to be addressed. -- Chip Salzenberg at Teltronics/TCT <chip@tct.com>, <uunet!pdn!tct!chip> "You can call Usenet a democracy if you want to. You can call it a totalitarian dictatorship run by space aliens and the ghost of Elvis. It doesn't matter either way." -- Dave Mack
sef@kithrup.COM (Sean Eric Fagan) (06/19/91)
In article <a4Nq417w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >No, the purpose of the news transport software is to correctly transmit the >news articles which it receives. The articles in question were not news articles. They did not meet the test for news articles (RFC's of various sorts), therefore, if C News did not transmit them, it was not failing its purpose. But I wouldn't expect you to understand or agree. -- Sean Eric Fagan | "I made the universe, but please don't blame me for it; sef@kithrup.COM | I had a bellyache at the time." -----------------+ -- The Turtle (Stephen King, _It_) Any opinions expressed are my own, and generally unpopular with others.
kucharsk@solbourne.com (William Kucharski) (06/19/91)
In article <a4Nq417w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >> Its purpose is to transmit *correct* articles. It succeeds. >No, the purpose of the news transport software is to correctly transmit the >news articles which it receives. > >C News failed in that purpose. No, this is like saying a compiler's purpose is to compile what I give it, not what's a legal program. Or, if you'd like, I can be nit-picky and say that if it doesn't obey the RFC syntax, it's NOT a news article, no matter what your posting software may say. -- | William Kucharski, Solbourne Computer, Inc. | Opinions expressed above | Internet: kucharsk@Solbourne.COM | are MINE alone, not those | uucp: ...!{boulder,sun,uunet}!stan!kucharsk | of Solbourne... | Snail Mail: 1900 Pike Road, Longmont, CO 80501 | "It's Night 9 With D2 Dave!"
mathew@mantis.co.uk (Giving C News a *HUG*) (06/19/91)
chip@tct.com (Chip Salzenberg) writes: > According to mathew@mantis.co.uk (Giving C News a *HUG*): > >The Problem is that there is buggy software. The C News > >silent-article-dropping does nothing to solve the problem, it merely hides > >the problem from the rest of the net. > > The word is not "hide", it's "protect." As in, "C News protects me > and my neighbors from bad articles that arrive at my site." Or in other words, "C News hides bad articles from my site". Same difference. Note that you completely ignored the actual POINT of the comment, which is that C News does nothing to SOLVE the problem of buggy posting software. > >I'm suggesting error-reporting because that would actually treat the cause o > >the problem, rather than the symptoms. > > Hardly. It enlarges the set of people who are notified about the > problem. But E-Mail notices can be ignored just as easily as log > entries. Hahah! Cunning. You conveniently fail to mention that the log entries you are talking about are on some other machine half way across the planet. Even if we were talking about log entries vs. email on the same machine, I don't think you would really believe that email was as easy to ignore accidentally. > Kent's concern about a local flood of bad news generating a global > flood of error messages may need to be addressed. The worst possible case is that we get one short error report for each bad article. This can easily be improved on; for example, we can make the error reporting program keep a log of sites it has issued an error for, and clear the log after (say) a week. So at most, we then get one error report per source site of a bad article per week. The recent Fidonet barf would have been no problem under such a scheme. mathew
mrl@uai.com (Mark R. Ludwig) (06/19/91)
In article <a4Nq417w164w@mantis.co.uk>, mathew@mantis writes: >chip@tct.com (Chip Salzenberg) writes: >> Its purpose is to transmit *correct* articles. It succeeds. > >No, the purpose of the news transport software is to correctly transmit the >news articles which it receives. Finally, the gauntlet is on the floor. This is the crux. Get out your biggest guns, folks (or longest RFC section numbers, as it were). This is where we stand or fall, because if we have no consensus about the *purpose* of the transport software, said software cannot succeed in its purpose. I agree with Chip's definition: its purpose is to transmit *correct* articles, as defined by the RFCs. Therefore, it must *not* transmit incorrect articles.$$ -- INET: mrl@uai.com UUCP: uunet!uaisun4!mrl PSTN: +1 213 822 4422 USPS: 7740 West Manchester Boulevard, Suite 208, Playa del Rey, CA 90293 WANT: Succinct, insightful statement to occupy this space. Inquire within.
tg@utstat.uucp (Tom Glinos) (06/20/91)
Mathew, give us a break. If you really feel strongly on the issue then do what Henry and Geoff did. (1) Think about the problem. Think about it HARD (2) Formulate a solution for it. (3) NOW IMPLEMENT IT. Goto (1) till satisfied and correct. (4) Get support for it. (5) Get it adopted. Unless you've completed steps (1)-(3), your comments hold little effective weight. When you are ready for steps (4) and (5) then you should use every possible means of pursuasion you can muster. Until then, give it a rest.
herrickd@iccgcc.decnet.ab.com (06/20/91)
In article <Bu2o43w164w@mantis.co.uk>, mathew@mantis.co.uk (Giving C News a *HUG*) writes: > > Firstly, is any effort being expended on fixing the problem? Henry seems to > have gone silent on the entire issue. > Yes, I just read a note that someone has put your From: in his kill file. dan
mathew@mantis.co.uk (Giving C News a *HUG*) (06/20/91)
sef@kithrup.COM (Sean Eric Fagan) writes: > In article <a4Nq417w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * > >No, the purpose of the news transport software is to correctly transmit the > >news articles which it receives. > > The articles in question were not news articles. They did not meet the test > for news articles (RFC's of various sorts), therefore, if C News did not > transmit them, it was not failing its purpose. Well, if you want to take that attitude, then from a pedantic viewpoint I suppose you're right. However, I maintain that it's not a very helpful attitude. It's much better to say "This is a news article, but it's badly formatted". Compilers don't just say "This isn't a C program" and silently delete the input; they give some useful information to the person who can correct the error. Even if you decide that articles not strictly RFC-conformant are not news articles, that does not remove all responsibility for reporting errors when such articles are encountered. I don't care how excessively pedantic C News is, so long as it reports its errors to someone who can do something about it. > But I wouldn't expect you to understand or agree. Of course you wouldn't, since you probably aren't listening to what I'm saying. mathew
mrm@sceard.Sceard.COM (M.R.Murphy) (06/21/91)
In article <DNJs414w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: [...] > >The worst possible case is that we get one short error report for each bad >article. > Your "we" is a bit centric, isn't it? The worst case is that every site that gets the error reporting newsgroup gets the report. If the error is reported through the generation of a new message, then perhaps even more sites get the report. It is a waste of a global resource to solve a local problem. Your solution is inefficient and not well thought. Questions aside: Is the result of dropping mathew's news articles silently equivalent to silently placing mathew in a kill file? Has his effective distribution coverage increased or decreased? -- Mike Murphy mrm@Sceard.COM ucsd!sceard!mrm +1 619 598 5874
scott@skypod.guild.org (Scott Campbell) (06/22/91)
In article <V4ku42w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >sef@kithrup.COM (Sean Eric Fagan) writes: >> The articles in question were not news articles. They did not meet the test >> for news articles (RFC's of various sorts), therefore, if C News did not >> transmit them, it was not failing its purpose. > >Well, if you want to take that attitude, then from a pedantic viewpoint I >suppose you're right. However, I maintain that it's not a very helpful >attitude. It's much better to say "This is a news article, but it's badly >formatted". Compilers don't just say "This isn't a C program" and silently >delete the input; they give some useful information to the person who can >correct the error. Sure, but if you then send a copy of that program to someone else and it doesn't compile on their machine, their compiler doesn't send you mail saying: "Sorry, Mathew, that construct isn't allowed on a VAX". YOUR compiler is the one that gives you error messages. Same for news. YOUR news posting program should tell you if your news article is correct or not, NOT mine. If you can't be bothered to follow the standards, and you can't be bothered to run news software written by someone who at least knows someone who has a friend who read an RFC or two, don't complain to the rest of the net that your postings are not getting out. Me? I don't need posting software, I have a copy of the RFC at my side and post by: echo '---my article--' | inews :-) scott -- Scott J.M. Campbell scott@skypod.guild.org Skypod Communications Inc. ..!gatech!dscatl!daysinns!skypod!scott 1001 Bay Street, Suite 1210 ..!uunet!utai!lsuc!becker!skypod!scott Toronto, Ont. (416) 924-4059 ..!epas.utoronto.ca!nyama!skypod!scott
cos@chaos.cs.brandeis.edu (Ofer Inbar) (06/23/91)
kucharsk@solbourne.com (William Kucharski) writes: >In article <a4Nq417w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: > >> Its purpose is to transmit *correct* articles. It succeeds. > >No, the purpose of the news transport software is to correctly transmit the > >news articles which it receives. > > > >C News failed in that purpose. > >No, this is like saying a compiler's purpose is to compile what I give it, >not what's a legal program. Or, if you'd like, I can be nit-picky and say >that if it doesn't obey the RFC syntax, it's NOT a news article, no matter >what your posting software may say. Now, I don't want to give the impression that I support one side or the other in this flame war (I don't know the issues involved, I missed the beginning of the thread), BUT: If you give a compiler something that does not meet the language spec, the compiler will generate an error message, and tell you what was wrong. So the analogy here is somewhat flawed. -- Cos (Ofer Inbar) -- cos@chaos.cs.brandeis.edu -- WBRS (BRiS) -- WBRS@binah.cc.brandeis.edu WBRS@brandeis.bitnet Broken Newsreader (TM): a newsreader which lacks certain features that my newsreader has, if said features would defuse my opponents' argument.
pa@verano.UUCP (Pierre Asselin) (06/24/91)
(Why am I jumping into this?) In article <1991Jun21.152743.22927@sceard.Sceard.COM> mrm@Sceard.COM (M.R.Murphy) writes: >In article <DNJs414w164w@mantis.co.uk> > mathew@mantis.co.uk (Giving C News a *HUG*) writes: >[...] >> >>The worst possible case is that we get one short error report for each bad >>article. > >Your "we" is a bit centric, isn't it? The worst case is that every site that >gets the error reporting newsgroup gets the report. If the error is reported >through the generation of a new message, then perhaps even more sites get the >report. It is a waste of a global resource to solve a local problem. Your >solution is inefficient and not well thought. Mathew's proposal isn't that bad. I can think of horror scenarios that cause local problems temporarily, but nothing that would clog the net globally. Mike may have something, though. Let's see... 1) A site posts an ill-headed message to an obscure but world-wide newsgroup. Instead, an error report goes out to news.errors; many more sites receive it than would have seen the original. 2) A site can reach the net at large through Cnews sites, but also through a Bnews-only path. If a message has non-compliant headers, Bnews sites worldwide see both the original and the error report. Potential increase of 100% in articles from that site, probably much less in bytes. Number (1) could be a problem for me. I get a limited feed. I can't pay for error stubs originating in alt.everything. Frankly, I don't know if it'll be a problem. I don't know how much traffic news.errors would carry. If the traffic is light, there is no problem. If the traffic is heavy, news.errors becomes ineffective and you might as well tell your feed(s) not to forward it to you. Of course, All the code writing, testing, patching and voting will have been in vain. I too would like to see some sort of error reporting mechan- ism. I expect it to be incomplete, unreliable and slow. General guidelines: - The pain should go where the problems are, because that's the only way they'll get fixed. If Mathew's messages are getting dropped, he and no one else should take care of it --but he has to know. - Any feature is a bug if you can't turn it off. The error-reporting mechanism had better be damned easy to switch off on the fly if anything goes wrong. - It has to be through the news. I don't see how mail- based notification has a prayer. A recent post to ca.test showed me that =my= mail feed is f*** up, possi- bly out of my control. I run B-news, so I'm a potential problem site. I don't want error notices bouncing between mailer daemons trying to reach me. The pain should go to me, not to a postmaster at a random site. Okay. Possible fix to potential problem (1). Someone tell me if this'll work. Modify Mathew's algorithm as follows: 5.5 Take the Message-ID of the bad article and replace the opening "<" with "<error". Add a new header, "Control: error". Create an error report article with this new Message-ID and control header. 5.6 Write an error report in the error article, and post the article to the original newsgroup. The RFC for the message-ID rewrites can cover the new control message as well. Meanwhile,, WE MUST BE SURE THAT EVERY PIECE OF NEWS SOFTWARE OUT THERE CAN SURVIVE AN UNRECOGNIZED CONTROL MESSAGE. Seems elementary, but you never know. I glanced through my news sources, Bnews would make an entry in the log file and mail a note to the newsadmin. The text might even end up in newsgroup control. I don't know enough to actively recommend anything. Only float ideas. If I have hereby demonstrated my ignorance of the news propagation and control mechanisms, flame me. Be brief. Be precise. -- --Pierre Asselin, Santa Barbara, California
mathew@mantis.co.uk (Giving C News a *HUG*) (06/24/91)
mrm@sceard.Sceard.COM (M.R.Murphy) writes: > In article <DNJs414w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * > [...] > >The worst possible case is that we get one short error report for each bad > >article. > > Your "we" is a bit centric, isn't it? The worst case is that every site that > gets the error reporting newsgroup gets the report. Yes. "We" being "the population of Usenet". > If the error is reported > through the generation of a new message, It isn't. It's reported through replacing a bad message with a good one. > It is a waste of a global resource to solve a local problem. News sites barfing badly-formatted news articles onto the net is not a "local problem". It affects everyone. mathew
mrm@sceard.Sceard.COM (M.R.Murphy) (06/25/91)
In article <81Z2445w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a *HUG*) writes: >mrm@sceard.Sceard.COM (M.R.Murphy) writes: >> In article <DNJs414w164w@mantis.co.uk> mathew@mantis.co.uk (Giving C News a * >> [...] >> >The worst possible case is that we get one short error report for each bad >> >article. >> >> Your "we" is a bit centric, isn't it? The worst case is that every site that >> gets the error reporting newsgroup gets the report. > >Yes. "We" being "the population of Usenet". In the case that your "We" is not just your site, then you are almost correct. The whole population of Usenet may receive the error report that is interesting only to your site. That is what I mean by a global solution to a local problem. It is Not A Good Thing. > >> If the error is reported >> through the generation of a new message, > >It isn't. It's reported through replacing a bad message with a good one. The good one is a new message. It isn't the same set of headers as the old message. The site that fed you the bad message will get the new message, too. BTW, you'll also have to change the Path: header, or none of the sites in the path will get the message. If you change the Path: header to contain just your (the posting site), I would suggest to you that qualifies as a new message. > >> It is a waste of a global resource to solve a local problem. > >News sites barfing badly-formatted news articles onto the net is not a "local >problem". It affects everyone. That's why C news drops the articles :-). It might be better if C news were to 1) Save the header or the whole offending dropped article, take your pick based on space available for the job. A reasonable way might be to append the headers to a file like /usr/lib/news/dropped. 2) In the course of normal daily cleanup, stripheaders <dropped | mail "$NEWSMASTER" cp dropped dropped.o cp /dev/null dropped 3) Have the script that does the deed for 1) be named ${NEWSBIN}/inject/mathew so that it can live near ${NEWSBIN}/inject/anne.jones 4) Apply :-) to 3) above. 5) The news administrator, being deluged with mail rather than having to remember to scan the logs, is more likely to let a site know that articles are being dropped. 6) If the site being dropped is not responsive, it wouldn't be much trouble to have stripheaders ignore that site or members of a list of sites. Or posters :-) The script stripheaders can be twisted from ${NEWSBIN}/inject/tear with very little effort. -- Mike Murphy mrm@Sceard.COM ucsd!sceard!mrm +1 619 598 5874
mathew@mantis.co.uk (Giving C News a *HUG*) (06/26/91)
scott@skypod.guild.org (Scott Campbell) writes: > Sure, but if you then send a copy of that program to someone else and it > doesn't compile on their machine, their compiler doesn't send you mail saying > "Sorry, Mathew, that construct isn't allowed on a VAX". YOUR compiler is > the one that gives you error messages. Yes, but equally the compiler doesn't prevent other people from trying to compile that program. I have no objection to news software refusing to send error reports to the source site, so long as it doesn't prevent the article from being received by other sites. mathew