RE: Schemas and Semantics (was: A Personal Reply...)

[Gavin Thomas Nicol <gtn@ebt.com>]

> If element type "foo" has specific semantics (and it 
> always does IMO), this doesn't mean that these semantics
> are obvious from the instance.

I understand your POV. My contention is that it is a very
limited view... because it is limiting the interpretation to
a single application domain where the interpretation is fixed.

For a good example of what I mean, look at HTML. When an
author writes HTML, they use the constructs for

  1) structural semantics
  2) formatting semantics

  -or-
   
  3) a combination of both

and you cannot know, a priori, what the intent was No
schema language that I am aware of today would make that
explicit.

FWIW. I have seen some environments where a single HTML
file passes through 1->3->2 when being served.