is the uniqueness constraint on top level components sufficient?

In the current draft, we say:

      <p> For each Binding component in the {bindings} property of a
      definitions container, the combination of {name} and {target
      namespace} properties must be unique.  </p>

(And similarly for Interface and Service components, the 3 top-level
components within <definitions> that are of concern for this issue.)

Thus, we only require that the QName of an interface, a binding or
a service be unique WITHIN the definitions container they're in. 
That is, we don't say its not ok for two <definition>s for the same 
target namespace to define two different interfaces and call give
them both @name="foo".

I think that's not sufficient - we should say that the QName MUST 
be unique, period. That is, its no koshure to define two bindings
with the same QName, for example. 

Yes I realize totally that there's no way to enforce that. However,
without such a constraint referring by QName, from a binding to an 
interface say, is not guaranteed to be correct. 

Thoughts? Can we please track this with an issue #? 



Received on Friday, 19 September 2003 04:59:37 UTC