>I find interesting and effective this alternative micro-document approach,
>don't understand what mechanism tells xml browser that the file names in
><file> have to be interpreted as file references (i.e. as files to
>not as any text strings.
At the moment there is no mechanism to tell the browser to interpret
elements as references. You will be able to express the required
relationship with XLink but XLink is still cooking...
Right now, I do it with "home grown" code. Instead of invoking
an XML parser directly I use two Python utilities with micro-documents:-
microparse.py -- parse document and any file's referenced from this file
macroparse.py -- parse document ignoring files referenced from this file
(same as a normal parse but can produce useful debugging output that
is specific to micro-document architecture approach).
Developers Day co-Chair WWW9, April 2000, Amsterdam