RE: ISSUE-19: Advice against using the \'all\' model group

Hey Paul!

> authors should think long and hard about using 1.0 all groups in the 
> schemas simply because their expressivity is so limited...not because 
> doing so will cause problems with any binding tool...right?

that's goes to the heart of the issue. Tools in general have little
problem consuming 'all'. It's the restriction on how schema authors
may then extend and use 'all' groups where the pain lies.

Paul

-----Original Message-----
From: Paul.V.Biron@kp.org [mailto:Paul.V.Biron@kp.org]
Sent: Wed 2/22/2006 9:49 PM
To: Downey,P,Paul,CXMA C
Cc: public-xsd-databinding@w3.org; public-xsd-databinding-request@w3.org; sandygao@ca.ibm.com
Subject: RE: ISSUE-19: Advice against using the \'all\' model group
 
> I think the thrust is not to preclude aspects of schema, just
> to advise on what works well with tools. So my preference would
> be to cite 'sequence' and 'choice' in the Basic documents and
> not mention 'all' at all, rather than include wording on why
> it should be avoided.

I'm curious...are there data binding tools that have known bugs/problems 
with all groups?  The limitations that schema 1.0 places on all groups 
should be something that all schema authors inform themselves about...but 
I can't see a databinding-specific reason to do that.  Rather, schema 
authors should think long and hard about using 1.0 all groups in the 
schemas simply because their expressivity is so limited...not because 
doing so will cause problems with any binding tool...right?

pvb

Received on Wednesday, 22 February 2006 23:35:35 UTC