RE: Impact of XML on Data Modeling

Thanks for pointing out those two bad practices of message design using
XML!

I would be very interested in some examples (if readily available), as
well as some general guidelines (good practices) in message design
(based on a common data model).
 
Please contact me directly if you prefer.
 
 
Thanks,

Scott Tsao, ATF
Architecture and Information Management
Boeing Information Technology
425-865-6676 MC 7R-09
Homepage: http://infoarchpre.web.boeing.com/People/Scott_Tsao/
<http://infoarchpre.web.boeing.com/People/Scott_Tsao/> 
Weblog: http://stsaoxml.blog.boeing.com
<http://stsaoxml.blog.boeing.com/>  


-----Original Message-----
From: Michael Kay [mailto:mike@saxonica.com]
Sent: Wednesday, January 30, 2008 2:43 AM
To: 'Pete Cordell'; Tsao, Scott
Cc: xmlschema-dev@w3.org
Subject: RE: Impact of XML on Data Modeling

> In fact, if you are just talking messages, there may be no defined
> 'application' data model at all (at least not an XML based one). 
> Protocols such as VoIP and HTTP are generally defined mainly in terms
> of the messages that flow between nodes.

Sure (though VoIP and HTTP are lower down the stack: they are channels
of communication rather than application message formats. Neither of
them knows anything about business objects).

I've certainly come across situations where people were struggling to
define messages between two applications because they hadn't defined the
common data model first. One application was using the term "retailer"
to refer to a shop, the other was using it to refer to a chain of shops
under common ownership, and untold confusion arose as a result.

The other side of the coin is that I've seen people try to define the
application data model as an XML Schema, and then wonder why their
messages were so large - they were sending masses of data that the
recipient didn't need, simply because the schema said it was mandatory.
They hadn't separated data modelling from message design.

Michael Kay
http://www.saxonica.com/