Re: Back-references and USEAGE

From: Usenet News Support (support@deathstar.prodigy.com)
Date: Tue Jun 29 2004 - 14:54:31 CDT


On Mon, 28 Jun 2004, Bruce Lilly wrote:

> Frank Ellermann wrote:
> > Bruce Lilly wrote:
> >
> >
> >>Given the fact that the Subject field is unstructured
> >
> >
> > RfC 1036 clearly defines a structure:
>
> RFC 1036 has no authority to [re]define the Structure field. It is
> defined in RFC 822 (and its predecessors and successor), and was
> specifically adopted by RFCs 850 and 1036 to replace "Title". And
> both 850 and 1036 clearly state that in cases of conflict, 822 is
> deemed to be authoritative and 850 and 1036 are deemed to be in error.
> So even if 1036 had attempted to define structure, such a definition
> would be null and void.

You keep repeating "unstructured" over and over, like an unsuccessful
magician repeating his magic word, and ignoring the problem that defacto
use is anything but unstructured. No matter what appears in a standard, in
the real world there is meaning to "Re:," "cmsg," part numbers in
multipart binaries, and in some cases "was" as well.

If you want to claim that this behaviour is a surprise to you, go to
it. Otherwise please apply your effort to solving the problem of how to
describe the world as it is, messy and somewhat inconsistent, in such a
way as to be useful as a guide to users and implementors regarding the
expected and possible behaviours of existing news software.

I'm not suggesting that these be required behaviours, just that the
document(s) produced by this WG not describe some well-behaved fantasy
world. It doesn't move this discussion forward to keep implying that we
can ignore this ugly reality by saying it's non-standard because the
Subject is unstructured.

What we need is a description of "this isn't required, but don't be
surprised if doing X results in Y," with Y being desirable or not by user
determination.

-- 
bill davidsen
  SBC/Prodigy Yorktown Heights NY data center
  Project Leader, USENET news
  http://newsgroups.news.prodigy.com
  Please honor the Reply-To header if present



This archive was generated by hypermail 2.1.7.