RE: simple case of IRIs for Components in WSDL 2.0

On Mon, 2005-09-12 at 11:34 -0700, Jonathan Marsh wrote:
> Are you suggesting that the form of the component identifier should be
> dependent upon what other (local) identifiers are in the document?  That
> we should define an identifier to a component that might lose its
> property of unique identification when other (unrelated but perfectly
> legal) components are added to the document?

Yes.

> This might be possible when you're trusting some infrastructure like
> Schemas, DTDs or xml:id to ensure no duplicate identifiers occur, but in
> our case duplicates (between symbol spaces) are completely legal.

I'd recommend that people don't use the same name for different
things in the same WSDL document; at least: not if they want
nice URIs for them.

> Your suggestion seems quite unstable in the face of WSDL evolution and
> the distribution of components between multiple documents, the full
> combination of which might not be available until runtime.

I'm aware of those risks and I think they're manageable.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Monday, 12 September 2005 18:41:38 UTC