Re: WWW meta indexes (proposal)
Tony Sanders <sanders@bsdi.com>
Errors-To: sanders@bsdi.com
Errors-To: sanders@bsdi.com
Message-id: <9310260421.AA03392@austin.BSDI.COM>
To: www-talk@nxoc01.cern.ch
Subject: Re: WWW meta indexes (proposal)
In-Reply-To: Tony Sanders's message of Mon, 25 Oct 93 13:57:10 CDT.
Errors-To: sanders@bsdi.com
Reply-To: sanders@bsdi.com
Organization: Berkeley Software Design, Inc.
Date: Mon, 25 Oct 1993 23:21:02 -0500
From: Tony Sanders <sanders@bsdi.com>
BTW: I found a pointer to the spec that my proposal was based on. It's
the IAFA (Internet Anonymous Ftp Archives) work which you can read about on:
ftp://sunsite.unc.edu/pub/docs/iafa/
namely:
ftp://sunsite.unc.edu/pub/docs/iafa/draft.part.II
I'll work through it and update the proposal (though on a first pass
I didn't see much that needs to be changed). I did add a Fax: field
on request from Terry Allen <terry@ora.com>.
I also changed Policy: to allow a URL using the <URL> notation.
And I added a Frequency: (per the draft) which is the "preferred" frequency
(in days) that index information should be updated.
Frequency: 5
I suppose, smart browsers/gateways could also use this information to aid
in setting a caching strategy (when more specific information is not
provided).
The full IAFA draft covers all kinds of information (abstracts,
packages, news and mailing list archives, etc). Once we have a
list of servers in place, we can worry about the other stuff.
We just need to keep in it mind.
The IAFA draft split SITEINFO (physical information) and DESCRIPT (logical
ifnormation) into two seperate files. They use filenames like IAFA-SITEINFO
which I think is a very unix centric thing to do. I choose site.idx
because it will work everywhere. I have no problem splitting this info
into two files if everyone thinks it's ok.
--sanders