xml-dev - Re: Real needs XML Schema must address (was re: W3C XML Schema Questiona

[Rick JELLIFFE <ricko@geotempo.com>]

Michael Champion wrote:
> 
> ----- Original Message -----
> 
> Rick Jelliffe wrote:
> 
> > One can evaluate XML Schemas on general terms as a universal schema
> > language. But it is more important at this stage, IMHO, to evaluate it
> > in terms of its sufficiency for meeting the pressing needs of the day as
> > the bottom line.
> 
> Absolutely.  Let's enumerate and discuss them.
> 
> I think everyone agrees that an XML Schema spec must:
> 
> - Allow schemas to be defined in XML syntax
> - Support the functionality of DTDs, more or less
> - Improve the data typing beyond what DTDs support
> - Allow namespace-aware validation
> 
> (dissent?)
 
I think that is where things get into a waffly hole: defining the the
requirements for XML Schemas in terms of technologies rather than domain
requirements. If XML Schemas supports scientific data well but not
ecommerce adequately, is that OK; if XML Schemas supports ecommerce well
but not XHTML, is that OK; if XML Schemas supports XHTML well but not
SOAP is that OK?  

There has been no definitive study made listing what people do with DTDs
(and I include my book, not to say that it isnt a good start): I mean a
study exhaustively grouping the various uses of parameter entities and
marked sections and external entities into abstract classifications. 
XML Schemas make a brave step forwards in this area.

Look at UML: they have 9 or so diagram types to cover common OO design
aproaches or portions.  And one couldn't say they are complete (they are
extensible, however); nor are they necessarily readily mappable from
non-OO techniques (i.e., figuring out how to draw DFDs in UML is not
obvious).  The same may be true of XML Schemas: we may find the future
of XML Schemas is to have it made of many "little" languages which model
only parts of the whole picture:
 * storage-class-based typing
 * grammars
 * key and unqueness constraints
 * etc

I, for one, would be very suspicious of any new schema language which
exactly fitted my personal requirements for today's projects with no
slack; if it seemed to be a little bit too big, but in some area I could
see would be useful to someone, I would have more confidence in it.

Rick Jelliffe

***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************