self-referencing Web indexes/documents

Kevin Altis (altis@ibeam.jf.intel.com)
Fri, 26 Nov 1993 13:49:13 -0800


Indexes to information on the Web are starting to pop up all over the
place. Until we have URNs and URN servers, most users will likely go to
NCSA for the NCSA Starting Points document, CERN for the complete Web
server list, etc.

I would like to reduce network traffic by keeping local copies of at least
the main index pages on our local server. Once a day or week via a cron
job, our local server could fetch the latest HTML document from the main
index points around the Web. I assume some people are already doing this.
Our server home page would then refer to the local copies of the indexes,
rather than the originals.

What's missing is a link within the HTML documents themselves, back to
their original location. This would allow the "local caching" to be
extended further to individual client machines and gives a workable
self-reference until HTML+ is fully implemented. With just regular HTML,
document authors could change their documents, to look something like:

<TITLE>Starting Points for Internet Exploration</TITLE>
<H1><A HREF
="http://www.ncsa.uiuc.edu/SDG/Software/Mosaic/StartingPoints/NetworkStartin
gPoints.html">Starting Points for Internet Exploration</A></H1>

That way, if someone wants to get the latest version to save to disk - this
is really handy for portable and home users browsing through documents
before connecting to the net - they just click on the link, get the latest
version and save to disk.

I know this will change once we have URNs and real distributed files, but
for now, there is usually a real home for a document, and I would like an
easy way to get it.

ka