Re: How to redefine <group> and <attributeGroup>?

Hi Eric,

> The text is very short and concise, for instance for groups:
[snip]
> I wonder it it's sufficient to describe group extension.

Hmm... yes, good point. I think the main problem is that the "schema
component constraint" is actually talking about the XML representation
of the schema rather than the schema components themselves. It would
probably be clearer if it talked about the particles that are allowed
as children of the redefining model groups. Taking inspiration from
the constraint given to prevent circular model groups:

  Within the {particles} of the group there must be at some depth a
  particle whose {term} is the group itself. There must not be more
  than one such particle. The {min occurs} and {max occurs} of that
  particle must be 1, as must the {min occurs} and {max occurs} of the
  model group that contains that particle and all its ancestor model
  groups [couldn't think of a neat way to say that].

Alternatively, it would be more restrictive but possibly easier to
define and implement if the redefined group was described as being a
valid extension of the original group.

Either of these would fix the other problem that you posted concerning
the redefinition of recursive structures because if you look at the
component level the element declaration is a black box.

Cheers,

Jeni

---
Jeni Tennison
http://www.jenitennison.com/

Received on Tuesday, 13 November 2001 15:10:54 UTC