Re: Is the concept of Element Equivalence Classes really needed?

"David Rosenborg" <david.rosenborg@omgroup.com> writes:

> Hi,
> 
> I think it would be more intuitive if you could specify an occurence
> of a class of elements by specifying just the allowed type, like in:
> 
> <type name="Foo">
>   ...
> </type>
> 
> <element name="bar">
>   <type>
>     <element type="Foo"/>
>   </type>
> </element>
> 
> <element name="foo" type="Foo"/>
> <element name="baz" type="Foo"/>
> 
> In this (experimental) example both foo and baz would be allowed as
> a child to bar.
> 
> Why do we need equivalence classes to accomplish this? To me
> they just add an unnecessary level of abstraction. What am I missing?

We had a design very close to this for a long time, but removed it
because of a number of weaknesses both easy to explain (you may
declare both 'shippingaddress' and 'billingaddress' to be elements
with the 'Address' type, but that does NOT mean that they are
interchangeable) and harder to explain (interactions with multiple
schemas and ambiguity rules).

The discussions are in the IG archive, which I'm sorry to say are
member-only.

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

Received on Wednesday, 19 January 2000 12:18:38 UTC