Ronald.Khoo@ibmpcug.CO.UK (06/25/89)
In article <1989Jun20.213129.8095@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes: >In article <10931@ibmpcug.UUCP> Ronald.Khoo@ibmpcug.co.uk writes: >>Another question: Why does only the max field in active get expanded to >> 10 digits but not the min field? (I'm obviously missing >> something obvious here, but I can't find it in the >> docs/READMEs) > >This is more or less an oversight. We consider the min field to be a >silly kludge for broken news readers, and pay very little attention to it. >So we missed that one. Uh.. I must'a missed something here - You recommend lwall's rn don't you? So do I - and I'm sure that it does use the min field - at least it seems to notice if I change a few by hand :-) If you have time, I wouldn't mind a quick two-line explanation of how you would replace the min field's function. Not a flame, Henry - I appreciate your many and varied contributions to the net, and now especially a nice lean & mean news transport - I'm just confused - and I don't want to install new software while I'm confused, but I'll kill for that fast expire! Thanks for listening.
henry@utzoo.uucp (Henry Spencer) (06/26/89)
In article <10947@ibmpcug.UUCP> Ronald.Khoo@ibmpcug.CO.UK writes: >Uh.. I must'a missed something here - You recommend lwall's rn don't you? >So do I - and I'm sure that it does use the min field - at least it seems >to notice if I change a few by hand :-) On the other hand, it doesn't do anything unpleasant if you just leave them all at zero. We do. >If you have time, I wouldn't mind a quick two-line explanation of how you >would replace the min field's function. The basic problem with the min field is that it doesn't supply enough information. The implicit assumption is that the min..max range is densely populated with articles, so no further information need be supplied. This just isn't so, given the possibility of articles with far-in-future explicit expiration dates. Any sensible news reader has to be prepared to go in and read the directory itself to find out what articles are really present. Given that, what use is min? -- NASA is to spaceflight as the | Henry Spencer at U of Toronto Zoology US government is to freedom. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
amanda@intercon.uu.net (Amanda Walker) (06/26/89)
In article <1989Jun25.174800.13276@utzoo.uucp>, henry@utzoo.uucp (Henry Spencer) writes: > Any sensible news reader has > to be prepared to go in and read the directory itself to find out what > articles are really present. Given that, what use is min? Well, for one thing, not every newsreader has access to the directory itself, which makes this approach a little difficult (NNTP clients, for example). Now, there may be better ways to figure out what's going on :-), but the min field at least gives the newsreader an idea of where to start checking for articles. Even in a sparsely populated group, earlier articles tend to expire before later ones... How much of an improvement to expire's speed does ignoring the min field really gain? -- Amanda Walker <amanda@intercon.uu.net> InterCon Systems Corporation -- "Those preachers are right--there's more to these songs than meets the eye..." --Arlo Guthrie
henry@utzoo.uucp (Henry Spencer) (06/28/89)
In article <26-Jun-89.101532@192.41.214.2> amanda@intercon.uu.net (Amanda Walker) writes: >... not every newsreader has access to the directory >itself, which makes this approach a little difficult (NNTP clients, for >example)... We class that as a deficiency in NNTP, best remedied by fixing the protocol. >How much of an improvement to expire's speed does ignoring the min field >really gain? It probably wouldn't make a whole lot of difference in speed. There are many useless features that could be added without much cost in speed. -- NASA is to spaceflight as the | Henry Spencer at U of Toronto Zoology US government is to freedom. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
lmb@vicom.com (Larry Blair) (06/29/89)
In article <1989Jun25.174800.13276@utzoo.uucp>, henry@utzoo.uucp (Henry Spencer) writes: > Any sensible news reader has > to be prepared to go in and read the directory itself to find out what > articles are really present. Given that, what use is min? Apparently the newly posted NN newsreader isn't sensible. It maintains an article database on which a crude expire is performed, based on the min field. A full rebuild of the database is too expensive to do every night. -- Larry Blair ames!vsi1!lmb lmb@vicom.com
amanda@intercon.uu.net (Amanda Walker) (07/03/89)
In article <1812@ucsd.EDU>, brian@ucsd.EDU (Brian Kantor) writes: > At the time we designed the protocol, no news reader needed a list of > all article numbers ahead of time. To my knowledge, still none do. > - Brian Well, actually, I wasn't complaining about a deficiency in NNTP or any specific problem with C news; I was objecting (very mildly...) to Henry's claim that newsreaders should expect to have to go rummaging around in the spool directory to find out what's there. My only point was that not all readers have that luxury. I'm quite happy with NNTP, actually. The only time I can think of that a newsreader could possibly want a list off all articles in a group is when it sees a newsgroup for the absolutely first time. -- Amanda Walker <amanda@intercon.uu.net> InterCon Systems Corporation -- "Those preachers are right--there's more in these songs than meets the eye..." --Arlo Guthrie
brian@ucsd.EDU (Brian Kantor) (08/08/89)
What deficiency? NNTP will a) give you the next article, and b) list and article either by article-id or by ordinal number in the current newsgroup. Oh, you wanted a list of valid article numbers ahead of time? You can get them the same way you would if you couldn't directly read the directory; given a minimum number and a maximum number, you could (if you really wanted to) step through and stat the corresponding filename. NNTP will let you do that if you wish. At the time we designed the protocol, no news reader needed a list of all article numbers ahead of time. To my knowledge, still none do. - Brian
dell@amelia.nas.nasa.gov (Thomas E. Dell) (08/27/89)
In article <1812@ucsd.EDU> brian@ucsd.edu (Brian Kantor) writes: >What deficiency? NNTP will a) give you the next article, and b) list >and article either by article-id or by ordinal number in the current >newsgroup. > >Oh, you wanted a list of valid article numbers ahead of time? You can >get them the same way you would if you couldn't directly read the >directory; given a minimum number and a maximum number, you could (if >you really wanted to) step through and stat the corresponding filename. >NNTP will let you do that if you wish. > >At the time we designed the protocol, no news reader needed a list of >all article numbers ahead of time. To my knowledge, still none do. > - Brian One of Rn's problems is that it does indeed have significant performance loss when skipping through large amounts of nonexistant articles. This is also true for Rrn, I've noticed. I've written something that uses opendir() etc to get minimum/maximum number (hilo) of messages. Unfortunately I have not extended this because of the NNTP problem. It is VERY useful for a newsreader to have information about valid article numbers, and I hope someone with influence over NNTP's direction will realize there is a demand for this feature. - T0M (dell@ames-nas.arpa)