As I see it, Kent's archive provides these advantages:
1) full text search,
2) active URLs,
3) splitting sections into separate pages (future).
No one else is providing all these in one archive. A couple of archives
provide full text search (and possibly active URLs), but not split
sections. OSU provides active URLs and split sections but not full text
search. rtfm.mit.edu provides the text but none of the above.
As has been pointed out, too, SmartPages is totally out of date and
should be purged. OSU is up to date but unsupported and could die on a
moment's notice. Actually without notice. The full text search archives
are not in the US and thus their use may cause delays and significant
added traffic. (Of course, the same argument applies to mirroring
Kent's archive in Europe.)
Thus I see split sections as a major contribution of a new archive. The
other two contributions are also valuable, but without split sections I
find it difficult to justify YAFA.
Yet ... many readers want to read an FAQ straight through and would
rather use the page down key on their keyboards rather than clicking
and waiting for a download for the next topic.
For this reason I think it's important to provide both forms, the full
article and the separate pages.
Here's a bunch of other comments in not much of any particular order.
For multipart FAQs, I'd love to see the parts combined with a single
index. I have an intro section and an index to all parts at the
beginning of each part; I'd like to see those elided. I can't say how
you should do that ...
You should handle at least all the faqmailfaqmail and structured formats that
Tom Fine handled.
Tom Lane mentioned displaying the Summary. I have the same Summary in
all parts, though I'll consider changing it. I include a word or two in
the Subject line which gives a clue as to what the part contains, so
the existing display is fine for the diabetes FAQ.
You should suppress the article headers except the most critical (From,
Newsgroups, Date, Reply-To, Subject, Summary). The complete header is a
lot of garbage to wade through at the top. Move it somewhere else to be
displayed on request. Some links in the headers should not be:
Originator could send useless email to faqserv@bloom-picayune.MIT.EDU;
the Supersedes: link is useless by definition; the Message-ID link is
of questionable value at best.
Tom Fine has a way to cross-reference sections within an FAQ. I've never
tried to make it work because it won't work across parts. If you define
a format that looks OK to human eyes and also converts to a link and
works across parts, I'd try to use the facility.
Link marking is a bit too aggressive for my taste. In addition to my
comments about the headers, I see far too many links to my own
newsgroup; it's a bit distracting to read a paragraph with
misc.health.diabetes underlined three or four times. A comment about
"the alt.support hierarchy" gets a news: link. I try to explain a valid
email address as "user@domain.typ"; this gets a mailto: link.
The search returns most hits as being in an .../index.html file. This
seems to be the file created when the FAQ is not split up. This is
confusing.
The search also shows raw html when it finds the target there. This is
very ugly and will confuse users, though I don't have a sound suggested
alternative.
Edward Reid
[
Usenet Hypertext FAQ Archive |
Search Mail Archive |
Authors |
Usenet
]
[
1993 |
1994 |
1995 |
1996 |
1997
]
© Copyright The Landfield Group, 1997
All rights reserved