OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Are we losing out because of grammars?

Going offlist to save some mush brained 
mail and the embarassment of not knowing 
what the heck this thread is about.
I think I agree, but I may be losing out 
on interpreting the score card of who 
wants what where.  Brain fry perhaps 
from writing an article about 
applying ontology-beyond-search.  It 
was a hard paper and I probably screwed it up.
Oh well. Back to the doofusBranch for Len...

In some cases, one wants to define a 
production (content model) that generates a 
context for a value.  In other cases, one wants to 
define a rule that tests a value in a context.  
The means for both have overlapping 
powers to express this, but one 
can be terser.  That is, a 
kolmogorov complexity test would suggest 
the superiority of one of these means for 
some output. In other cases, one and only 
one of these means works.

I understand that there are things 
which can't be checked from the 
schema grammar (co-occurence) and 
that is fine so far.  It is a real 
and immediate requirement to apply 
languages such as Schematron in addition 
to an XML Schema.

I can't understand if someone wants to:

1) Get rid of schema and use rules-based 
languages in all cases or vice versa

2) Make sure that both rules and grammars 
are expressible in schemas *ASAP* thus tossing 
XML Schema back to committee for another year.

I guess I understand that one might want 
to ensure 2) can be done at some future 
point by some standard means such as what 
Clark and the others are discussing.  I 
understand the need to do both. Given 
a document en route among different owners 
using different systems with different local 
rules I understand the need to be able to 
separate into different documents rules 
that encode the production of contexts from 
documents that check values in contexts.

So I think I understand what is lost if 
only grammars are provided.  But given 
Schematron, that isn't the case.  So where 
is this thread headed?


Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h