NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/comp/infosyst/gopher/2067

Internet Message Format | 1993-01-25 | 2.6 KB

Path: sparky!uunet!mcsun!uknet!yorkohm!nigelm
From: [e-mail] (Nigel Metheringham)
Newsgroups: comp.infosystems.gopher
Subject: Re: Gopherising FAQ files
Message-ID: <[e-mail]>
Date: 25 Jan 93 18:33:35 GMT
References: <[e-mail]>
Organization: Electronics Department, University of York, UK
Lines: 50

[Following up to my own posts - maybe I should just talk to myself]

In my previous article I wrote about the "gopher-ising" of Usenet
FAQ files.  This would mean that on a gopher server you would see a
list of FAQ directories, and in each directory you would see a set
of entries each being one section (question/answer?) of an FAQ file.

Additionally there would be an index for each FAQ set, and ideally
an index for all the FAQs from the top level.

Basically since FAQ files are free format we have to "persuade"
gopher and WAIS to recognise their subsectioning.  Actually all you
*need* to do is get them into a format that WAIS understands - you
can then put together a set of gopher links (using range addressing)
from the WAIS index.cat file with a simple piece of Perl.

FAQ files are mostly different from one another.  They also do not
fit neatly into any of the known WAIS document types.  It would be
quite easy to co-erce them into a pseudo mail-digest form, but this
doesn't really present the right information (the From: field would
tend to be redundant).

So of the types already handled by WAIS, it would seem best to munge
FAQs into a dash separated format, where each section is separated
by a line of dashes.   The first line following the dashes is the
section identifier.  This leaves the FAQ still readable, but able to
be directly indexed by WAIS.  A script can be written to modify most
FAQ sets into this format quite easily.

the disadvantage is that the FAQs need to be rewritten (ie sepearate
file store).

I have also, despite misplaced confidence, been unable to do
anything similar with NeXT indexing (ie I had to go and build WAIS
instead).  The NeXT indexer seems unable to cope with ranges within
a file even if you replace the commands it uses to analyse a file
(because the rotton thing stats the file before starting anything
else, and it objects if you give it a gopher range type filename).


I would appreciate it if anyone with strong feelings as to the
"right" way to split FAQs would contact me about it before I do too
much work on this technique!

    Nigel.

--
#   Nigel Metheringham  -- (NeXT) EMail: [e-mail]    #
#   System Administrator, Electronics Dept, University of York    #
#   York YO1 5DD. Phone: [phone], Fax: [phone]      #