How about: <http://www.isogen.com/papers/archintro.html>, *A Tutorial
Introduction to SGML Architectures*
I can't disagree that the name is a poor one (the name was chosen before I
got involved with the HyTime standard) and that the AFDR annex is not a
very accessible document, for which I must take responsibility as the
HyTime editor primarily responsible for managing the text. I hope the
document above goes a little way towards undoing the damage. We (ISOGEN)
also provide a one-day SGML Architectures course that we've taught at
several conferences to what I think is reasonable success (everyone seemed
to get the concepts).
I think one of the problems with the architecture mechanism is that, like
SGML, the basic idea is very simple but the simplicity is obscurred by the
details of the mechanism, which is complicated by the need for automapping
to make the use of architectures convenient for document authors. I think
David Megginson's approach with architecture support on top of SAX does a
good job of cutting down to the bare minimum you need to do something
useful (however, there are some important things that you may need to do
that you can't with David's implementation, in particular, multiple levels
of architectural derivation).
Cheers,
E.
-- <Address HyTime=bibloc> W. Eliot Kimber, Senior Consulting SGML Engineer ISOGEN International Corp. 2200 N. Lamar St., Suite 230, Dallas, TX 75202. 214.953.0004 www.isogen.com </Address>