rick@seismo.UUCP (03/13/87)
Currently, 2.11 news goes to great pains to keep unknown newsgroups on the Newsgroups line, but not localize them. This has caused many people to change their mind and claim that this is really wrong. A favorite example is somthing posted to misc.jobs,ut.jobs from ut-sally ending up in the "local" ut jobs group at U of Toronto. Does anyone still think that the unknown groups should be preserved or should a future patch remove the unknown newsgroups? ---rick
stephen@comp.lancs.ac.uk (Stephen J. Muir) (03/15/87)
In article <43152@beno.seismo.CSS.GOV> rick@seismo.CSS.GOV (Rick Adams) writes: >Does anyone still think that the unknown groups should be preserved or >should a future patch remove the unknown newsgroups? It should remove them. A more important reason is that, for example, although Europe doesn't get talk.* newsgroups, many sites think we do because of the cross-posted articles from other newsgroups we *do* get! Then these users post to a talk.* newsgroup and those articles get junked with no warning to the poster. Also, my /usr/lib/news/errlog keeps filling up with "unknown newsgroup talk.whatever not localized". This file should not grow quickly or news administrators will be discouraged from looking at it. -- EMAIL: stephen@comp.lancs.ac.uk | Post: University of Lancaster, UUCP: ...!mcvax!ukc!dcl-cs!stephen | Department of Computing, Phone: +44 524 65201 Ext. 4120 | Bailrigg, Lancaster, UK. Project:Alvey ECLIPSE Distribution | LA1 4YR
jbuck@epimass.UUCP (Joe Buck) (03/16/87)
Sorry this is so long, but I think a detailed discussion is needed. In article <43152@beno.seismo.CSS.GOV> rick@seismo.CSS.GOV (Rick Adams) writes: >Currently, 2.11 news goes to great pains to keep unknown newsgroups >on the Newsgroups line, but not localize them. This has caused many >people to change their mind and claim that this is really wrong. I think it's right. The standards document said this was how news was supposed to behave; 2.11 is the first one to get this right. It's now less critical for every non-leaf site to have a correct active file; either the article will be rejected (and possibly flow around the bad site) or the Newsgroups line will be left alone. For example, let's say site "hao" (for example) decides not to accept talk groups. Under 2.10.x, to avoid trashing the Newsgroups line on cross-posted articles, they'd have to maintain a line in active, and a directory, for all the talk groups. Local readers (including the manager who ordered "talk" to be cut off) could still read the talk groups, though only cross-posted articles would be present. The alternative (Unknown group talk.philosophy.misc removed) would be to ruin the Newsgroups line, since the article may eventually get to a part of the network where talk exists. Under the new regime, hao can simply remove the talk groups. No hassle. The sys admin doesn't have to keep track of "talk" group creations and deletions, and other sites don't get mad at them. >A favorite example is somthing posted to misc.jobs,ut.jobs from >ut-sally ending up in the "local" ut jobs group at U of Toronto. Hmm. Well, nothing's perfect. The problem is that there's no way to say that the two ut.jobs groups are different. What if the two groups were "sci.math,talk.philosophy.misc" (this has been a fairly common cross-posting at times)? Then this is exactly the behavior you want. Go back and read Mark Horton's document (in the doc directory with the 2.11 distribution). He explains this point. One solution that's sure to make everyone scream: make all distributions unique. But it's too late for that. >Does anyone still think that the unknown groups should be preserved or >should a future patch remove the unknown newsgroups? I think it's more critical than ever (with overloaded backbone sites dropping whole categories of groups) that the current 2.11 behavior be preserved. I think the new approach causes fewer problems than the old approach. Another point. Say you're a leaf site (or otherwise totally dependent on a single upstream site, though you may feed others) and the upstream administrator has a missing group. Under 2.10.x the upstream site strips names; you have no idea anything's wrong. Under 2.11 the name of the new group is there, in cross-postings. You get messages in the log file about it (unknown group not localized) so you can ask the upstream administrator what is going on. -- - Joe Buck {hplabs,ihnp4,sun,ames}!oliveb!epimass!jbuck seismo!epiwrl!epimass!jbuck {pesnta,tymix,apple}!epimass!jbuck Entropic Processing, Inc., Cupertino, California
rick@seismo.CSS.GOV (Rick Adams) (03/17/87)
> you want. Go back and read Mark Horton's document (in the doc > directory with the 2.11 distribution). He explains this point. Unfortunately Horton has reversed his position and favors removal. Thats what prompted the general question. (The current opinon is about 60% for removal and 40% for keeping. I was hoping for something more definitive) ---rick
grr@cbmvax.UUCP (03/17/87)
In article <43159@beno.seismo.CSS.GOV> rick@seismo.CSS.GOV (Rick Adams) writes: > >Unfortunately Horton has reversed his position and favors removal. Thats >what prompted the general question. > >(The current opinon is about 60% for removal and 40% for keeping. I was >hoping for something more definitive) I definitly favor keeping the groups - 1) it bridges gaps in distributions, allowing the articles to fall into their intended groups no matter how they arrive. this allows the reader to distinguish a cross posted article from an apparently mis-posted one. 2) it allows news to flow during name changes and similar confusions by simply putting both groups in the header. the current moribund state of mod.amiga.{sources,binaries} is a case in point. Perhaps some thought should be given to naming conventions for regional and local groups and stripping groups that being exported from a distribution class i.e. local, regional, net... -- George Robbins - now working for, uucp: {ihnp4|seismo|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@seismo.css.GOV Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
dhb@rayssd.RAY.COM (David H. Brierley) (03/17/87)
In article <297@dcl-csvax.comp.lancs.ac.uk> stephen@comp.lancs.ac.uk (Stephen J. Muir) writes: >In article <43152@beno.seismo.CSS.GOV> rick@seismo.CSS.GOV (Rick Adams) writes: >>Does anyone still think that the unknown groups should be preserved or >>should a future patch remove the unknown newsgroups? > >It should remove them. A more important reason is that, for example, although >Europe doesn't get talk.* newsgroups, many sites think we do because of the >cross-posted articles from other newsgroups we *do* get! Then these users post >to a talk.* newsgroup and those articles get junked with no warning to the >poster. To get the most functionality, I think the software should do both. The unknown group should be preserved in the copy of the message that is sent on to other sites but should be removed in the copy that is stored locally. This way, the other sites can still see the cross posting if they want to but local users are not confused by the strange newsgroups name. Unfortunately, I realize that this would probably require some extensive changes to rnews/inews since the copy passed along to other machines would now be different than the one stored locally. >Also, my /usr/lib/news/errlog keeps filling up with "unknown newsgroup >talk.whatever not localized". This file should not grow quickly or news >administrators will be discouraged from looking at it. This doesn't seem like a real big problem to me. I have an awk program that runs every night that analyzes the log file and sends me a summary via email. If my site did not support the talk. groups I would just add a line to the awk script that recognized "unknown newsgroup talk." and threw it away. It doesn't take much effort to write an awk script to analyze the log file, you can probably even use the ones Erik Fair wrote (they have been posted to the net a couple of times). -- David H. Brierley Raytheon Submarine Signal Division; Portsmouth RI; (401)-847-8000 x4073 smart mailer or arpanet: dhb@rayssd.ray.com old dumb mailer or uucp: {cbosgd,gatech,ihnp4,linus!raybed2} !rayssd!dhb
eppstein@tom.columbia.edu (David Eppstein) (03/17/87)
Obviously some groups in some contexts (talk in na) should be kept in the message regardless of whether the link wants that group. Obviously other groups and contexts (the two different uts outside their respective universities) would be better stripped. So, the obvious solution is to make both possible on a group-by-group basis. Why has no one else already proposed such a solution? -- David Eppstein, eppstein@cs.columbia.edu, Columbia U. Computer Science Dept.
piet@mcvax.cwi.nl (Piet Beertema) (03/17/87)
>Does anyone still think that the unknown groups should be preserved or >should a future patch remove the unknown newsgroups? Remove 'em. I see no reason to preserve newsgroups that aren't received locally (or that aren't even received in e.g. Europe). On the contrary: it's utterly confusing if such newsgroups are preserved. That's also exactly the reason why we worldwide checkgroups control messages are unwanted: large parts of the world do *not* receive all newsgroups. -- Piet Beertema, CWI, Amsterdam (piet@cwi.nl or mcvax!piet)
heiby@mcdchg.UUCP (Ron Heiby) (03/18/87)
I also favor the current 2.11 behaviour of maintaining the newsgroups in the header as posted (or, of course, modified by aliases) for the reasons given by several others in this forum. In article <297@dcl-csvax.comp.lancs.ac.uk> stephen@comp.lancs.ac.uk (Stephen J. Muir) writes: >Also, my /usr/lib/news/errlog keeps filling up with "unknown newsgroup >talk.whatever not localized". This file should not grow quickly or news >administrators will be discouraged from looking at it. The way I deal with this, is the following code in the .profile for the netnews administrator login. (Note that my LIBDIR is not in the default location.) Maybe I'm trimming too much out of errlog. This seems to get rid of the trash. ----- if [ -s $HOME/lib/errlog ] then cat $HOME/lib/errlog | egrep -v 'Unknown newsgroup|Newsgroups in active|Duplicate|Aliased|Orphaned|unopenable' > $HOME/lib/errlog.new mv $HOME/lib/errlog.new $HOME/lib/errlog fi if [ -s $HOME/lib/errlog ] then echo "Stuff in $HOME/lib/errlog!" # if there's still something fi ----- -- Ron Heiby, mcdchg!heiby Moderator: mod.newprod & mod.os.unix Motorola Microcomputer Division (MCD), Schaumburg, IL "Save your energy. Save yourselves. Avoid the planet 'cuae2' at all costs!"
henry@utzoo.UUCP (Henry Spencer) (03/18/87)
The C news implementors (Geoff and I) favor retaining unknown groups, due
to a fundamental belief that the right thing to do with headers is to leave
them alone. (The need to compromise this for Path and Xrefs is annoying.)
The problem of name clashes between regional groups is indeed troublesome,
but just stripping out unknown groups is not a good solution.
--
"We must choose: the stars or Henry Spencer @ U of Toronto Zoology
the dust. Which shall it be?" {allegra,ihnp4,decvax,pyramid}!utzoo!henrydave@rsch.wisc.edu (Dave Cohrs) (03/19/87)
It seems that what we need is a way to tell inews what to localize
and what not to localize, without changing the headers, of course.
A quick (meaning I didn't spend hours thinking about this) solution
would be to add another control file which lists distributions and
the sites that can post to them. Something like:
comp all
soc all
uw *.wisc.edu
etc,etc
[ let's assume such a file lives on uwvax, aka rsch.wisc.edu ]
This way, if uwvax receives a message for "misc.forsale,uw.general", it
would look to see which site originated the message. If it isn't in the
list of "localizable" sites, it doesn't localize the distribution.
Comments?
Dave Cohrs Proud member of NOTHING
+1 608 262-2196 UW-Madison Computer Sciences Dept.
dave@rsch.wisc.edu ...!{harvard,ihnp4,seismo,rutgers}!uwvax!davejgd@uwmcsd1.UUCP (03/22/87)
Numerous previous articles have made a case for zapping "unknown" newsgroups
from the Newsgroups: header line.
[Lighting a match... ]
Much as I hate to do this [ :-) ], I will argue against this change by citing
the *documented STANDARD*. (It's a dirty trick, but someone has to do it!)
==> RFC 850 June 1983
==> Standard for Interchange of USENET Messages
==> Mark R. Horton
==>
==>
==> [ This memo is distributed as an RFC only to make this
==> information easily accessible to researchers in the ARPA
==> community. It does not specify an Internet standard. ]
==>
==> 1. Introduction
==>
==> This document defines the standard format for interchange
==> of Network News articles among USENET sites. It describes
==> the format for articles themselves, and gives partial
==> standards for transmission of news. ...
==>
==> [Non-germane text deleted]
==>
==> 2.1.5 Newsgroups The Newsgroups line specifies which
==> newsgroup or newsgroups the article belongs in. ...
==>
==> If an article is received with a Newsgroups line listing
==> some valid newsgroups and some invalid newsgroups, a site
==> should not remove invalid newsgroups from the list.
==> Instead, the invalid newsgroups should be ignored. For
==> example, suppose site A subscribes to the classes
==> "btl.all" and "net.all", and exchanges news articles
==> with site B, which subscribes to "net.all" but not
==> "btl.all". Suppose A receives an article with
==> "Newsgroups: net.micro,btl.general". This article is
==> passed on to B because B receives net.micro, but B does
==> not receive btl.general. A must leave the Newsgroup line
==> unchanged. If it were to remove "btl.general", the
==> edited header could eventually reenter the "btl.all"
==> class, resulting in an article that is not shown to users
==> subscribing to "btl.general". Also, followups from
==> outside "btl.all" would not be shown to such users.
Now, although the above cited RFC does not purport to be an Internet
standard, it *does* claim to reflect USENET standards. I submit that
News B.2.11 conforms to this standard (at least insofar as section
2.1.5 is concerned.)
Recent proposals are to "break" News 2.11. (By removing "unknown" newsgroups
from the "Newsgroups: header.)
If people want to change the way News works, they should change the
standards first. (Or at least *propose* changing the standards, *THEN*
change the software.) [This is a *religious* position -- don't argue! :-)]
Please keep in mind that as more and more sites start dropping branches of
the news directory tree, the situation described in the example of 2.1.5
will become more common. If we start dropping "unknown" newsgroups from
the headers, (more?) little "black holes" will start appearing in USENET.
Enough articles get dropped on the floor as it is. Besides, "junk" will
start getting more activity, and who *really* wants to read "junk"?
[Extinguishing match... just spotted Smokey the Bear.]
--
John G Dobnick
Computing Services Division @ University of Wisconsin - Milwaukee
UUCP: {ihnp4|uwvax|uwmacc}!uwmcsd1!jgd
INTERNET: jgd@csd1.milw.wisc.edu
"Knowing how things work is the basis for appreciation,
and is thus a source of civilized delight." -- William Safire