🚀 go-pugleaf

doug@cc.ysu.edu (Doug Sewell) writes:
>There's a real storm brewing over in news.admin.misc (primarily) in the
>thread "Re: Do something about missing References: headers".

Dear Doug (or other news.admin.misc readers),

I tried to go read that thread, but the start of it has fallen off my
server already and there're still tons of messages.

Could someone summarize why References: should be enforced? I never use
them, but I'd like to understand why people think it is so important as
to risk destroying bit.* to insist on it.

Thanks,
	Ian Chai
	BERITA-L (bit.listserv.berita) listowner

In article <38rela$voq@unix1.cc.ysu.edu>, doug@cc.ysu.edu (Doug Sewell) says:
>Basically what's being proposed is trashing  "non-rfc-compliant"
>articles.  In theory, it could make much of bit.* go away, and also
>e-mail gateways into other newsgroups as well (the inet groups, for
>example).

I understand the concern about broken threads, but the fact is that
the information is simply unavailable to the gateway. I see three
possible resolutions:

(1) Apply the check at POST time instead of for in-stream posts. This
    would, in effect, leave things as they are now for the gateways.
(2) Apply the check only for hierarchies in the traditional ("Big 7")
    hierarchies. All mailing lists gated to these groups would have to
    be dropped, but hierarchies such as bit.** would be unaffected.
(3) Have the gateway remove the "Re:" on the subject line. These 
    are not replies in the news sense of the word, but rather new posts.

Of these three options, the first would be best.

Jim McIntosh (jim@american.edu)
The American University
Washington DC 20016-8019 USA

In article <38RLU3$56P@PALADIN.AMERICAN.EDU>, JIM@american.edu writes:
> I understand the concern about broken threads, but the fact is that
> the information is simply unavailable to the gateway. I see three
> possible resolutions:
>
> (1) Apply the check at POST time instead of for in-stream posts. This
>     would, in effect, leave things as they are now for the gateways.

This is in any case the real solution.  Newsreaders (clients in general) have
to be RFC-compliant, as much as the servers ought to be compliant.   Now,  it
is already known that NETNEWS is not correct on all parts of the RFC 1036, so
one could argue that the client can be a bit more liberal also in that  case,
but I doubt anyone will deny that this is a very false conclusion.   At  post
time, all needed information is available (when it is a follow-up), so  there
isn't much reason not to follow the standard.

> (2) Apply the check only for hierarchies in the traditional ("Big 7")
>     hierarchies. All mailing lists gated to these groups would have to
>     be dropped, but hierarchies such as bit.** would be unaffected.

This is not possible, since the RFC clearly handles about ALL newsgroups,  no	
matter whether they are amongst the 'standard' USENET hierarchies or not.

> (3) Have the gateway remove the "Re:" on the subject line. These
>     are not replies in the news sense of the word, but rather new posts.

Ughh, no please.  That would be a last resort-action, and even then I guess a
rewrite of 'Re: foobarbaz' to 'Was: foobarbaz' would be the  only  acceptable
way of doing that, since there ought to remain some reference to the original
posting.

Kris.