Re: is the uniqueness constraint on top level components sufficient?

Sanjiva, 

it seems to me that the restriction you're trying to add to WSDL would
be better-suited (and actually enforcable) in the sources of WSDL
components. For other WSDL components, the spec already solves this.

Your example before has indicated that an entity knows an Interface by
its QName, looks it up somehow (in a registry of sorts) and uses it. I
think it's the registry that should specify that there is only one
Interface for one QName.

I think the status-quo is good.

                   Jacek Kopecky

                   Senior Architect
                   Systinet Corporation
                   http://www.systinet.com/



On Fri, 2003-09-19 at 04:59, Sanjiva Weerawarana wrote:
> "Martin Gudgin" <mgudgin@microsoft.com> writes:
> > 
> > The spec currently says ( for interfaces, for example )
> > 
> > For each interface component in the {interfaces} property of a
> > definitions container the combination of {name} and {target namespace}
> > properties must be unique. 
> > 
> > There is only ever one definitions container. Therefore there is only
> > one {interfaces} property. Therefore interfaces MUST all have unique
> > Qnames. If they don't it's an error and the WSDL parser should
> > catch-fire-and-die. 
> 
> Agreed; WSDL parsers are not the ones who have the problem I'm
> trying to solve.
> 
> > I don't understand what it is you want to change.
> 
> The wording in the spec does not guarantee that there MUST only
> be exactly one Interface component with a given QName. 
> 
> Maybe I'm just being retarted and not seeing that there's no
> problem, but what I had in mind was to add a statement like the
> following somewhere:
> 
>     "The QName identifying any top level Interface, Binding or 
>      Service component MUST be unique from amongst all similar
>      components. That is, two different top level components of
>      the same kind MUST NOT share the same QName. However, two
>      components of different kinds MAY have the same QName."
> 
> Sanjiva.
> 
>

Received on Monday, 22 September 2003 13:06:32 UTC