timk@wynnds.xenitec.on.ca (Tim Kuehn) (05/19/91)
In <1991May18.211109.20401@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes: >In <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.) What are these obvious schemes - besides sending an email to the offending site for each badly formatted article? How about if a cron script tabulated the type of offenses Cnews found over a given period of time (say, a week) and posted a tabulation to control, news.lists, or some other group reporting the which sites are posting broken articles, what they're doing wrong, and how often they're doing it wrong? The report could consist of a header line detailing the code of the offenses reported, and then a list of detail lines with the name of the offending systems and what they're doing wrong. The general format would be: <hdr line detailing the offense codes reported> <detail lines with the offending system name and offense list> <detail lines with the offending system name and offense list> <detail lines with the offending system name and offense list> And would look like this: xx=offense1, yy=offense2, zz=offense3, etc. <site_name>: xx(999), yy(999), zz(999) This could potentially generate a fair bit of traffic on the target newsgroup (until people got their broken SW fixed), but it'd surely be less traffic than is found in some groups that are already in existance (like alt.desert-*) and a lot more useful in pointing out who's got the broken software on the net without seriously breaking or overloading the offending site's mail spool, or their upstream spool. I haven't seen this scenario in the list of 'obvious alternatives.' Comments? ------------------------------------------------------------------------ Tim Kuehn TDK Consulting Services (519)-888-0766 timk@wynnds.xenitec.on.ca -or- !{watmath|lsuc}!xenitec!wynnds!timk Valpo EE turned loose on unsuspecting world! News at 11! "You take it seriously when someone from a ballistics research lab calls you." Heard at a Unix user's meeting discussing connectivity issues.
henry@zoo.toronto.edu (Henry Spencer) (05/22/91)
In article <1991May19.035200.879@wynnds.xenitec.on.ca> timk@wynnds.xenitec.on.ca (Tim Kuehn) writes: >How about if a cron script tabulated the type of offenses Cnews found >over a given period of time (say, a week) and posted a tabulation to >control, news.lists, or some other group reporting the which sites are >posting broken articles, what they're doing wrong, and how often they're >doing it wrong? A reasonable idea for a *few* sites to do. (We did something along those lines before the software changes were released, in fact, although we did it by mail rather than by news.) Not so good if it's part of the standard configuration, because the last thing we need is a group with thousands of largely-redundant articles every week. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
timk@wynnds.xenitec.on.ca (Tim Kuehn) (05/22/91)
In article <1991May21.191941.15498@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes: >In article <1991May19.035200.879@wynnds.xenitec.on.ca> timk@wynnds.xenitec.on.ca (Tim Kuehn) writes: >>How about if a cron script tabulated the type of offenses Cnews found >>over a given period of time (say, a week) and posted a tabulation to >>control, news.lists, or some other group reporting the which sites are >>posting broken articles, what they're doing wrong, and how often they're >>doing it wrong? >A reasonable idea for a *few* sites to do. (We did something along those >lines before the software changes were released, in fact, although we >did it by mail rather than by news.) Not so good if it's part of >the standard configuration, because the last thing we need is a group >with thousands of largely-redundant articles every week. Ok then, if you need is a mechanism to keep notifications from being too excessive and redundant, limit posts of about sources of offending articles to systems that are *immediate neighbors*. For example, if site "x" fed site "y" and "y" detected bad articles POSTED from (not passing through) "x", it posts a *summary* of the faults found somewhere public, with something more detailed logged to an appropriate file. Since most sites talk to a resonably small number of other sites, posts of possible bad articles would be kept to a manageable number as opposed to having all kinds of machines post notes about bad articles being written. The remaining problem is keeping the distribution down to a manageable area so the whole world doesn't hear that your neighbor is posting bad articles. The obvious problem with this scenario is that sites who aren't neighbors to a latest-patchlevel-Cnews system wouldn't be informed they're doing a news no-no. I have an idea on that, but want to take some more time thinking about it. ------------------------------------------------------------------------ Tim Kuehn TDK Consulting Services (519)-888-0766 timk@wynnds.xenitec.on.ca -or- !{watmath|lsuc}!xenitec!wynnds!timk Valpo EE turned loose on unsuspecting world! News at 11! "You take it seriously when someone from a ballistics research lab calls you." Heard at a Unix user's meeting discussing connectivity issues.
) (05/22/91)
henry@zoo.toronto.edu (Henry Spencer) writes: > In article <1991May19.035200.879@wynnds.xenitec.on.ca> timk@wynnds.xenitec.on > >How about if a cron script tabulated the type of offenses Cnews found > >over a given period of time (say, a week) and posted a tabulation to > >control, news.lists, or some other group reporting the which sites are > >posting broken articles, what they're doing wrong, and how often they're > >doing it wrong? > > A reasonable idea for a *few* sites to do. How about if one or two sites do it, and mail the guilty parties, and in the mean time C News doesn't try to enforce its own idea of what the standards are? If after six months there are sites who are still simply ignoring the automatically-generated warnings, then perhaps we can consider more unpleasant methods. mathew
red@redpoll.neoucom.edu (Richard E. Depew) (05/23/91)
In article <1991May22.123413.5485@wynnds.xenitec.on.ca> timk@wynnds.xenitec.on.ca (Tim Kuehn) writes: >In article <1991May21.191941.15498@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes: >>In article <1991May19.035200.879@wynnds.xenitec.on.ca> timk@wynnds.xenitec.on.ca (Tim Kuehn) writes: >>>How about if a cron script tabulated the type of offenses Cnews found >>>over a given period of time (say, a week) and posted a tabulation to >>>control, news.lists, or some other group reporting the which sites are >>>posting broken articles, what they're doing wrong, and how often they're >>>doing it wrong? > >>A reasonable idea for a *few* sites to do. (We did something along those >>lines before the software changes were released, in fact, although we >>did it by mail rather than by news.) Not so good if it's part of >>the standard configuration, because the last thing we need is a group >>with thousands of largely-redundant articles every week. > >Ok then, if you need is a mechanism to keep notifications from being too >excessive and redundant, limit posts of about sources of offending articles >to systems that are *immediate neighbors*. For example, if site "x" fed >site "y" and "y" detected bad articles POSTED from (not passing through) >"x", it posts a *summary* of the faults found somewhere public, with >something more detailed logged to an appropriate file. > >Since most sites talk to a resonably small number of other sites, posts >of possible bad articles would be kept to a manageable number as opposed >to having all kinds of machines post notes about bad articles being written. >The remaining problem is keeping the distribution down to a manageable >area so the whole world doesn't hear that your neighbor is posting bad >articles. Isn't this a good use for the "to.neighbor" newsgroups? Each summary or list would then be posted on only two machines, the notifying site and the neighboring site. It could even be made a part of the standard configuration without danger of flooding the net with these warnings. Dick Depew -- Richard E. Depew uunet!aablue!{redpoll,neoucom}!red Munroe Falls, OH. red@redpoll.neoucom.edu (home) NEOUCOM, Rootstown, OH. red@uhura.neoucom.edu (work)
kucharsk@solbourne.com (William Kucharski) (05/25/91)
In article <71mc313w164w@mantis.co.uk> mathew@mantis.co.uk (CNEWS MUST DIE!) writes: >How about if one or two sites do it, and mail the guilty parties, and in the >mean time C News doesn't try to enforce its own idea of what the standards >are? The point is C News isn't enforcing its own idea of what the standards are, it's enforcing what the RFC's ideas are of what the date standard is. -- | 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!"