RE: DTD vs Schema

AMER, Nidal (Nidal.AMER@hdmp.com)
Fri, 11 Sep 1998 15:19:23 +0200

In fact I am trying to implement in XML an object model I have designed
with UML in which I clearly define generalization and specialization
relationships. In instance, I am treating a generic type called
transaction from which I derive different types of transactions all of
which share a common behavior.=20
The XML Data schema uses SUPERTYPE to declare inheritance, even if this
is not very useful unless the DOM model, built by the parser, could
recognize that an element inherits another one and lets the application
treat the subtype the same way as the supertype ( polymorphism ). Also,
when referenced, the subtype should be checked against applying the
validity constraints of the supertype plus its own constraints.

Unfortunately, as this seems not to be implemented, I have to build =
this
logic in the code.

Regards,
Nidal.

-----Original Message-----
From: Peter Murray-Rust [SMTP:peter@ursus.demon.co.uk]
Sent: Friday, September 11, 1998 2:21 PM
To: AMER, Nidal
Subject: Re: DTD vs Schema

At 12:44 11/09/98 +0200, you wrote:
>Thank you all a lot for your answers. I am a step further now.
>Just I still can't express inheritance with DTD in an elegant
way.

No - it's not easily built in to XML. I think that we shall
develop some
general means for doing this. It depends *what* your want to
inherit.

I have developed this in my Chemical Markup Language,
http://www.xml-cml.org (BTW I used to work in the pharma
industry). I have
an array element which can be subclassed to atomArray and
bondArray (and
all of these are mapped to Java). But there is no way of
*enforcing* the
parallel between XML and Java here - I have to remember it when
I write the
code.

I think DCD and other schemas will start addressing this.

P.

>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>Nidal AMER
>Deputy R & D Director
>Health Data Management Partners (HDMP)
>A SmithKline-Beecham Company
>6 Rue de Gen=E8ve,
>1140 Brussels, Belgium
>Tel: + 32 (2) 724 00 93
>Fax: + 32 (2) 726 91 59
>E-mail: nidal.amer@hdmp.com <mailto:nidal.amer@hdmp.com>=20
>Visit our web site: http://www.hdmp.com <http://www.hdmp.com>=20
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
Peter Murray-Rust, Director Virtual School of Molecular
Sciences, domestic
net connection
VSMS http://www.nottingham.ac.uk/vsms, Virtual Hyperglossary
http://www.venus.co.uk/vhg