From: Eivind Tagseth (eivindt@multinet.no)
Date: Sat May 01 2004 - 03:37:19 CDT
* Charles Lindsey <chl@clerew.man.ac.uk> [2004-04-30 17:09:37 +0000]:
> In <20040429145212.GA4859@tagseth-trd.consultit.no> Eivind Tagseth <eivindt@multinet.no> writes:
> >Now, if all newsreaders were compliant newsreaders, then you're right, any
> >such newsreader would not have to take into account any other "back-references"
> >than "Re: ". In fact, the back-reference wouldn't be of any use at all
> >(please correct me if I'm wrong).
>
> I think you are possibly confusing "compliant reading agents" (which we
> don't define anywhere) with "compliant followup agents" (which we do
> define).
Yes, I was. Sorry about that.
> If followup agents were allowed to prepend "Sv: ", then reading
> agents which tried to do anything other than pure References-based
> threading would have to take account of that (which could mean that they
> worked better in your example). But then there would also be hundreds of
> translations of "Re:" into Greek and Arabic and Chinese and whatever else.
> So implementors of reading agents would not know what they were expected
> to do. Currently, they mostly know what to do with "Re:".
Well, some obviously don't, or we wouldn't see all these "Sv: Re: Sv: RE: Sv: "
subjects. And until all followup-agents are compliant with this spec,
we'll continue seeing these, won't we? And as Seth said, with the current
wording, a followup-agent wouldn't even be allowed to suggest a subject
with these "back-references" truncated to a single "Re: ", so new compliant
followup-agents will continue being part of the problem (adding one extra
back-reference if the first back-reference wasn't "Re: "). A followup-agent
not adding back-references (also compliant given the current wording) would
at least not contribute _more_ back-references prefixed to the subject.
> So my problem with John's text is that it allowed for no distinction
> between "Re:" (which is commonly supported) and those hundreds of others
> (which are not).
John's position, as far as I understand it, is that there should be no
notion of a "back-reference" in the subject, as it is an unstructured
header. So how _could_ he support one over the others, that would make
no sense! I'm sure John will have more to say about "back-references"
when we get to USEAGE, where this discussion IMO belongs.
Eivind