I think a whole lot of folks have already thought a whole lot about
problems like this, and they thought this stuff was important enough
to standardize on.
They call their work HyTime.
We should take a look at it and see if we can't borrow from what they
did.
For example, it should be trivial to derive this LINK element from the
hytime ilink architectural form. See
Message-ID: <19930511.074728.98@almaden.ibm.com>
Date: 11 May 1993 09:36:53 -0400 (19930511.133653)
Newsgroups: comp.text.sgml,alt.hypertext
Subject: Re: WWW and Hytime (was Explicit Linking is Impossible)
at:
wais://ifi.uio.no/comp.text.sgml/TEXT/11048/1=ifi.uio.no%3A210;2=/local/ftp/lib/wais/comp.text.sgml;3=0%2011048%20/local/ftp/pub/SGML/comp.text.sgml/by.msgid/19930511.074728.98%40almaden.ibm.com;4=ifi.uio.no%3A210;5=/local/ftp/lib/wais/comp.text.sgml;6=0%2011048%20/local/ftp/pub/SGML/comp.text.sgml/by.msgid/19930511.074728.98%40almaden.ibm.com;7=%00;
for one way of doing it.
Dan