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

Yes, I concur that if we need a way to group thigs by *name* your
method is preferable over the superfluous equivalence class construct.
However, in a system where we have types I believe that the most natural
way of grouping is by type, not by names of instances. Infact, the behaviour
you propose could be achieved with my proposal as well:
If we want to create an "open" group we could do the following:

<group name="facet" order="choice">
  <element type="numFacet"/>
</group>

<element name="precision" type="numFacet"/>
<element name="scale" type="numFacet"/>

But in this case the group isn't really needed. It would be, if we wanted to
group
elements of heterogeneous types but still allow for bottom up extension:

<group name="FoosNBars" order="choice">
  <element type="Foo"/>
  <element type="Bar"/>
</group>

Cheers,

</David>





"Curt Arnold" <carnold@houston.rr.com> on 2000-01-22 06:14:00

Please respond to "Curt Arnold" <carnold@houston.rr.com>

To:   David Rosenborg/OMT/OMGROUP@OMGROUP
cc:

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


Your discussion is quite along the same lines as my earlier note

http://lists.w3.org/Archives/Public/www-xml-schema-comments/2000JanMar/0040.
html

I'm baffled what arguments could be used to justify the complexity of
equivClass's when the same thing can be accomplished much more simply with
groups.