W3C home > Mailing lists > Public > www-forms@w3.org > August 2000

Re: Doesn't XForms duplicate what already exists?

From: Goetz Bock <bock@blacknet.de>
Date: Tue, 29 Aug 2000 01:04:49 +0200
To: www-forms@w3.org
Message-ID: <20000829010449.A3300@titan.blacknet.de>

Hi Adam,

> [ ... ] a good portion of the XForms initiative is duplicating what 
> already exists. 
Kind of

> Take the data model for instance. From what I can tell, it is designed to
> describe the data structure of XML data. Isn't this exactly what XSchema and
> DTDs are supposed to do? 
Yep, but the "lightwight syntax" described in the "XForms 1.0: Data Model
W3C Working Draft 06 April 2000" should be a simplevied version for
people that don't have the time/will to lern DTD/XMLschema.

Or to use W3C wording:
"It uses XML Schema for the semantics of the data types, but re-expresses 
 this in a simple syntax more likely to win the hearts and minds of typical 
 HTML authors."

> [ ... ] Does it really make sense to write one document to
> describe valid data and another to describe how to pass valid data? The
> should be functionally the same thing. Perhaps XSchema doesn't provide
> enough robustness to do all the stuff XForms data model does but shouldn't
> that mean that XSchema needs to be changed?
While the validation is done in the data model, I'm not sure this is the
right place for it. Things like

<string name="spouse" required="status is 'married'"/>

should not be part of the data model (but that's just MHO).
Most (if not all) things required by XForms can be acived by a namespace
addition to XSD. 

> [ ... ] If we have a standard that already defines valid
> data, why is necessary to create a new standard. It seems that writing
> XSchema should be no different that writing an XForms data model and we
> should, in general, be writing XSchema (or DTD or whatever) for our XML data
> anyways, [ ... ]
BTW: It's possible (more or less) to use an XSLT to map (the old, and
only) XForms data model into XMLSchema and the other way should be 
possible as well.
 
Cu,
    Goetz.
Received on Monday, 28 August 2000 19:05:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 10 March 2012 06:21:48 GMT