Re: Radical cure for BOS confusion (3 CC's deleted)
At 8:57 AM 1/9/97, Derek Denny-Brown wrote:
>On a third note, in the recent past a number of people have hinted to moving
>some of the hyperlink info into the style-sheet (or a style-sheet). I just
>want to point out that not all hyperlinks necessarily exist for human
>traversal. I can think of a number of cases where I might be sending XML
>documents between (remote) processes and want to use hyperlinks to express
>information relationships, where this data is never directly presented to
I have not just hinted at it, I advocate it as a way to allow proper
processing of links by browsing applications, which are our prototypical
example of an application that does not need a DTD. And of course, with a
different notion of stylesheet, other forms of application processing than
browsing might also be accomodated without a DTD.
If you want to _analyse_ link relationships, then I think a requirement
to fetch the DTD is reasonable. What I really want is to use architectural
forms and default attribute values, but not require authors to put
defaultable attribute values in their instances for the benefit of
DTD-challenged applications. XML browsers will almost certainly require
stylesheets anyway, so that adding linking to the stylesheet allows link
behavior to be preserved, in just the case where _all_ you really care
about is behavior.
No application should have to analyze a turing-complete script to
determine if it invokes link behavior or not. There must always be some
_declarative_ way to determine if a random piece of markup is a link.
I am not a number. I am an undefined character.
David Durand email@example.com \ david@dynamicDiagrams.com
Boston University Computer Science \ Sr. Analyst
http://www.cs.bu.edu/students/grads/dgd/ \ Dynamic Diagrams
MAPA: mapping for the WWW \__________________________