From: Charles Lindsey (chl@clerew.man.ac.uk)
Date: Mon Aug 23 2004 - 10:31:03 CDT
In <4120DE65.8080800@epix.net> "Forrest J. Cavalier III" <mibsoft@epix.net> writes:
>Henry Spencer wrote:
>> On Mon, 16 Aug 2004, Forrest J. Cavalier III wrote:
>>
>>>>The problem with that is that part 1 of a FAQ is related to part 2...
>>>
>>>I understand USEFOR to document article format, so it isn't clear to me that
>>>considering your point is necessary for USEFOR.
>>
>>
>> The format includes what the headers *mean*, so it's quite correct to ask
>> when a References header is supposed to occur and what it is supposed to
>> point to.
>>
>I guess I see this as akin to a "class" vs "instance" discussion.
>We are defining what articles will include a References header. I do not agree
>we are defining what it is supposed to point to. Let me be clear that the
>content of the references header is not mentioned in the definition of followup
>which I have proposed.
No, you are wrong. USEFOR is Syntax + Semantics. So with any header,
including the References-header, it is right and proper to say:
a) What it looks like (syntax - in this case a list of message-ids);
b) Which articles MUST/SHOULD/MAY contain that header (followups, or
whatever else we choose to say);
c) What the header "means" - that is the part you seem to want to omit,
but it is essential for every header, as Henry has pointed out.
So what is the "meaning" of the References-header?
It is *advice* to certain reading agents (those that choose to make use of
it, since it is not obligatory) as to how to present/display that article.
In particular, the article in question should (l.c. "should") be
presented/displayed *after* all those listed in the References-header;
moreover, those listed later in the References-header should be
presented/displayed after those listed earlier.
No, that is not a wording ready to plug into our draft, but a wording
satisfying that aim could surely be written. Note that there is no such
wording in the present draft because there is a process described under
"Duties of a followup agent" which achieves the required effect, but if we
are going to use References-headers not generated by followup agents, then
some "meaning" wording ought to be written. Note that RFC 2822 also
describes the meaning of the References-header by giving a process for
constructing it.
>Further, I am having trouble thinking of what will be "wrong" on the reading
>agent when Part 1 has a References pointing to part 2. The author must have
>had some reason for doing it that way, and the reading agent is going to
>have a display order set by references which will probably match whatever
>the author intended. It is a good example to consider, but the answer need not
>change the definition I proposed for USEFOR.
By including References-headers in multipart FAQs, the author is giving
*advice* as to how he would like those FAQs presented/displayed (and in
general the order of such presentation/display may well be different from
the order of posting). It would indeed be bizarre for an author to ask for
Part 1 to be displayed after Part 2 (or more likely a bug in his posting
script).
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl Email: chl@clerew.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5