Dave Schweisguth <dcs@proton.chem.yale.edu> wrote:
>Andrew Main <zefram@dcs.warwick.ac.uk> wrote:
>> Another possibility is to avoid using lines that PGP will escape. It
>> won't then be possible to adhere to RFC 1153, but that isn't really
>> suited to FAQs anyway.
>
>The last statement is certainly not true, as the existence of any number of
>useful and readable FAQs in RFC1153 shows. Personally I find RFC1153 (which
>is how Tom Fine HTMLifies our FAQs, and which also serves as a convenient
>format for entry headings and date stamps) more useful than any sort of
>authentification, and won't bother with the latter until it permits the
>former.
Alright, my statement wasn't *quite* what I meant. RFC 1153 is not *ideally*
suited to FAQs, and in my opinion is not *well* suited to FAQs. It was
intended for the collation of loosely related articles from (possibly)
different authors, and as such works well. *Some* FAQs are adequately
organized in this way. But many FAQs are by a single author (or small number
of authors), and have some structure beyond a simple division into sections.
For multi-level structure, RFC 1153 produces rather awkward Subject: lines,
primarily because the section number must be treated as part of the subject
line. Furthermore, section headings that follow RFC 1153 look ugly.
The best solution (IMHO) is for a header line to tell the reader how to find
the subject headings. A simple regular expression would work, and wouldn't
limit FAQ authors in the matter of how to format. If this idea is taken
seriously enough to actually be implemented (unlikely), I suggest a header
Section-Regexp: in the auxiliary header (or maybe X-Section-Regexp: in the
posting header). Furthermore, if there is a multi-level structure, two or
more r.e.s could be specified, in Section-Regexp-1:, Section-Regexp-2:, etc.
- -1: would be the regexp for the lowest level. For example, for my FAQ:
Section-Regexp-1: /^[XVI][XVI]*\.[1-9][0-9]*\. /
Section-Regexp-2: /^[XVI][XVI]*\.1\. /
As an extension, I suggest that multiple regexps be allowed in each header,
which must match consecutive lines. This would allow for section headings that
are more difficult to confuse with normal text. Thus my headers would become:
Section-Regexp-1: /^$/^$/^[XVI][XVI]*\.[1-9][0-9]*\. /^$/
Section-Regexp-2: /^$/^$/^[XVI][XVI]*\.1\. /^$/
(Two blank lines before the heading -- which wouldn't appear anywhere else --
and one after.) If I remember the minimal digest format correctly (and I may
not), it would have
Section-Regexp: /^------------------------------/^$/^Subject: /
Comments, please?
On a separate matter:
Tomas Willis <tomas@cae.wisc.edu> wrote:
>By the way, I think we should all adopt "URL:" in the secondary header.
><ducking>
>(Really.)
I'd like to say "no" to that, but not without qualification. I think a
standard way to signal where:
1. the latest edition of the FAQ,
2. other formats of the FAQ
reside would be a Good Thing. However, URL: is too vague. It will also cease
to be a good idea when URNs come into use. To indicate the URL, use the
<URL:whatever> convention. I suggest headers such as Latest-Version-Plaintext:
and Latest-Version-HTML:, etc. As other formats become available (note
*available*, not *common*), they can each get a Latest-Version-whatever:
header line. The header for each FAQ should only list the formats that the FAQ
is actually available in, so this standard avoids the stumbling-block of
previous suggestions, because it does not require that the FAQ maintainer
support any particular format (even plaintext).
One final note: I think these extra header lines should go in the auxiliary
header, not the posting header. The posting header is for information about
posting to USENET, whereas the auxiliary header is for information about the
FAQ.
I'm sorry to re-open all these old arguments, but I do think I have new
suggestions here. (I couldn't see anything like this in the archive.)
-----BEGIN PGP SIGNATURE-----
Version: 2.6
iQBVAgUBLxXQh7UPMZ11+Ju9AQGAOgIA0a0O0P8bDSRedOlb1//qb8ji1IyxIJil
vYi/sfHMnZAr6PzQ1ZSbEbDQeu6Iw1ZnZig1pPLEhYJ7v5sZ+UFcSA==
=3LZc
-----END PGP SIGNATURE-----
-- Andrew Main <zefram@dcs.warwick.ac.uk>
[
Usenet Hypertext FAQ Archive |
Search Mail Archive |
Authors |
Usenet
]
[
1993 |
1994 |
1995 |
1996 |
1997
]
© Copyright The Landfield Group, 1997
All rights reserved