Re: targetResource wording

On Sun, Jun 15, 2003 at 08:15:07PM -0600, Champion, Mike wrote:
> > if it really can identify anything, then I believe my
> > previous concerns are relevent.  If it's only intended to identify "an
> > interface that encapsulates", then let's give it an appropriate name
> > that doesn't include the word "resource".
> 
> Whoa, back to Restifarian kindergarten for one of us!  In what sense is an
> interface -- some software object with identity, that is reachable by a Web
> protocol -- not a "resource?"

It *is* a resource.  But not all resources are interfaces.

(and that's a unique definition of "interface" you've got there! 8-)

> The point of the WSDL 1.2 targetResource proposal, as I understand it from
> the WSA perspective, is to let the application developer writing the WSDL
> file figure out whether it is referring to (ahem) a document, a service, a
> person, an organization, a physical object, or a concept...

Sanjiva just said that it identified a chunk of software (if I
interpreted him correctly) which implemented the interface.  DaveH told
me he thought it could identify any resource, such as a dog.  And you
appear to be saying it's for identifying types of resources?  Yikes!

From where I'm sitting, Sanjiva's definition seems to make the most
sense.  So in that case, we're back to naming; while it's true that the
chunk of software is a resource, it is *not* true that all resources
are chunks of software.  Therefore the attribute should not be called
"targetResource", as that name suggests that any resource identifier can
be a valid value.  How about just "target"?

MB
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca

Received on Sunday, 15 June 2003 23:59:40 UTC