From: Bruce Lilly (blilly@erols.com)
Date: Sat May 01 2004 - 14:50:47 CDT
Nick Boalch wrote:
> Bruce Lilly wrote:
>
>> Since you have not addressed any of the substantive issues, no.
>> It is absolutely mind-boggling that you state that you have
>> "tried to take their comments on board" when your proposed text
>> is littered with the term "back-reference", which both John and
>> I have pointed out is meaningless in the context of readers and
>> reading agents (i.e. as you have used it).
>
>
> I think perhaps you fail to understand the nature of compromise.
>
> Hopefully you will be able to accept that all of the following points
> are true:
>
> (1) Use of "Re: " to initialize the Subject-header of followups is
> widespread and current behaviour on Usenet.
Perhaps, but that has nothing to do with
a) defined syntax or semantics
b) network operations
and therefore does not belong in documents addressing those
topics. It might be appropriate for some other document.
> (1.1) When "Re: " is used in this way, it can be referred to as
> a 'back-reference'.
Possibly in the context of a followup agent, but not in the context
of a reading agent or (human) reader, which/who has no way of determining
whether any string is or is not a "back-reference", i.e. the term is
hopelessly ambiguous.
> (1.1.1) Other strings are sometimes seen being used as
> back-references; however, their use causes lots
> of problems for followup agents constructing
> Subject-headers according to current practice.
I believe you are confusing followup agents and reading agents. There
is no problem for followup agents.
Regarding reading agents, the problems with text other than re are no
greater than the problems caused by "Re:", viz. reading agents may
wish to strip such text (with the danger that it is stripping something
essential to the topic) when sorting by subject.
> (1.2) However, it is problematic for agents confronted with a
> given Subject-header to decide with 100% certainty what
> constitutes a back-reference and what doesn't.
They can do so with precisely *no* certainty, and that is one of the problems.
> (2) Subject-header-initial "Re: " is used by some reading agents to
> sort messages for display to readers.
I think you mean it is stripped (ignored), rather than "used".
> (2.1) A better way to sort messages for display is to use the
> References-header.
There is not necessarily a "better" or "worse" way in general. If one
is specifically speaking of threading, defined as using References
relationships, then the only way to do that correctly is to use the
References fields -- there is no substitute.
> Do you accept all of those points? If not, I don't believe we'll ever
> manage to come up with a wording that you won't nitpick somehow.
Do you disagree with the analysis point-by-point above? If so, let's
discuss.
> (1) Some people on this mailing list note that use of "Re: " is common
> and familiar to readers, and want to mandate its use in the
> initialization of the Subject-header of followups.
There is no basis for that in the defined syntax or semantics of the
Subject field. There is no basis for that regarding network operations.
It would be unduly prohibitive in terms of user agent freedom to do
sensible things (e.g. determine Subject field content based on AI
analysis of body text, or present the user with a dialog indicating that
the subject field constant should be entered manually).
> (2) Some people on this mailing list think that mandating use of "Re: "
> conflicts with the unstructured nature of the Subject-header.
They're correct.
> (3) Most people are fed up to the back teeth of the whole thing and
> desperately want to move on.
We can easily defer the matter to the third document which we are
supposed to produce. It is only the insistence of a few people (our
Document Editor being one of them) that is preventing us from moving
on.
> We do not appear able to reconcile viewpoints (1) and (2), so the ONLY
> COURSE OF ACTION LEFT OPEN TO US is to produce a text that steers a
> middle course.
Wrong. Why is it that you fail to see that the Standards Track documents
need say nothing at all about user interface issues?
> I think this is what the text I have proposed does. It explains the
> situation, covering all the points above, and leaves the decision about
> whether to prepend "Re: " automatically to the implementors of followup
> agents.
That has nothing to do with the topic matter appropriate for either of
the Standards Track documents which we are working on. It might be
appropriate for a "practices" document.
> My proposed text does not deprecate use of "Re: ", as you would prefer,
> because this WG has not been able to come to consensus about doing so.
> It does not mandate the use of "Re: ", as I would prefer, because this
> WG has not been able to come to consensus about doing so.
Irrelevant w.r.t. the Standards Track documents.
> It uses the
> term 'back-reference' because that is convenient.
No, it is worse than inconvenient, because it is meaningless in any
practical sense. Given the field:
Subject: Re: is an anachronism
is there a back-reference, and if so, what is it?
> The only thing that I find "absolutely mind-boggling" is your absolute
> refusal to alter your position one iota for the sake of getting off this
> issue and moving on.
The issue of what is and is not appropriate for the Standards Track
documents is not a matter of opinion; the topics have been clearly
specified by the WG chairs.
> I am beginning to get to the stage where I think we
> should just start to ignore you because you have stopped contributing
> anything useful and started being just an irritating, repetitive
> obstruction to progress.
If you cannot distinguish between specified syntax and semantics, and
a religious devotion to a 4-character string, perhaps you are the obstruction.