From: Eivind Tagseth (eivindt@multinet.no)
Date: Tue Aug 24 2004 - 02:11:30 CDT
* Charles Lindsey <chl@clerew.man.ac.uk> [2004-08-23 18:22:22 +0000]:
> In <20040817072224.GB27760@tagseth-trd.consultit.no> Eivind Tagseth <eivindt@multinet.no> writes:
>
> >We should just keep it clear and simple. Keep the definition of References
> >as it is today, keep the definition of followup as it is today (or make
> >it more clear), and move the multi-part-controversy into USEAGE --
> >e.g. make a section advising people to post multi-part messages (FAQs,
> >patch-sets, etc) as if they were followups in one discussion to simplify
> >finding, killing, scoring, etc.
>
> No, you can't do that, because using References for multi-part FAQs would
> then violate a MUST in the References-header (and you cannot advise people
> in USEAGE to do that). The minimal fix is to relax that MUST in the
> References-header (which is what I have proposed). The same effect can be
> achieved by altering the definition of 'followup'; some people seem to
> prefer that.
Of course you can do that, unless you have a really strict definition of a
followup. You want to have this strict definition, while relaxing the
definition of the protocol. This feels really backwards to me, as the
protocol is the part that must be really well defined in order to secure
interoperability, and is also the part that is most easily non-ambiguously
specified. Intent, type of agent, content of subjects and message bodies
are not elements of a precise, non-ambiguous, definiton.
Eivind