Re: wildcards

I believe that you are the second person within a few days who has asked 
for something that the workgroup considered as the "variant GI" model of 
substitutability. In other words, you wish to write a content model in 
which any element name (GI) can appear, as long as it has the expected 
type. I will give you essentially the same answer that I gave to him:

After months of agonizing debate, the schemas workgroup declined to 
support the variant GI model, at least in this version of this 
specification.  I do not think that I could speak for the group as a whole 
in trying to net out the reasons, and it is not clear that all of us would 
agree on either the nature or the merits of such reasons. These are very 
difficult and subtle (though useful!) facilities to introduce into a 
language, once you start to consider details. 

You might check out section 3.6.5 in the December 17th draft, which 
describes element equivalence classes. It is possible that they would meet 
your needs. 

------------------------------------------------------------------------
Noah Mendelsohn                                    Voice: 1-617-693-4036
Lotus Development Corp.                            Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------------







Imran Rashid <imranr@dragonfly.wolfram.com>
Sent by: www-xml-schema-comments-request@w3.org
12/21/99 03:54 PM

 
        To:     www-xml-schema-comments@w3.org
        cc:     (bcc: Noah Mendelsohn/CAM/Lotus)
        Subject:        wildcards


I'm not sure if the current definition of wildcards can do what I want or
not, but I don't think it can.

For some documents I'm writing, there is one place in the document where
the user could come up with any element they want.
however, the contents of that element are rigidly defined.  currently, to
make sure the documents are still always valid, I'm having to dynamically
generate a DTD, even though it's.  99.99% the same for each document.

so I was hoping that with schema's I could come up with a way of saying
that in one spot, any element can be used, but that element must follow
this content model.

in other words, could I do something like:

<element name="topElement">
  <type>
    <any>
       <group ref="myGroup"/>
    </any>
  </type>
</element>


so you could clarify the option/ exclusion of content-models like that in
the wildcard section.
(and if you're going to exclude it, I'd really like to see it added.)


thanks,
Imran Rashid

Wolfram Research, Inc.
www.wolfram.com

Received on Tuesday, 21 December 1999 17:22:32 UTC