Re: Fwd: Error in diagram?

David, I agree there's a problem with the current definition ;-), but 
I'm not quite sure I agree with your revised definition :-(

First, the definition should have read "a service represents" and not 
"an interface represents" (serious typo).

Second, I have an issue with using "other resource", which gives the 
impression there are two resources. Next, I think your second sentence 
breaks the flow with the one that follows in the spec, namely:

"A service can thus be seen as a collection of endpoints bound to the 
same resource."

Finally, I'm not quite sure why we can't use "Web resource", which I 
think was Arthur's original proposal. Can't a Web resource be a 
"physical object or something abstract or conceptual"?

To summaryze, I propose that we rewrite the whole paragraph as:

<current>
WSDL further considers that an interface represents (some of) the 
behavior of a resource on the World-Wide-Web. A service can thus be seen 
as a collection of endpoints bound to the same resource.
</current>

as follows:

<proposed>
WSDL further considers that a service may actually be providing access 
to a Web resource (note: a Web resource may be a physical object or 
something abstract or conceptual). A service can thus be seen as a 
collection of endpoints bound to the same resource. Multiple services 
may be related if they point to the same resource.
</proposed>

What do you think?

Jean-Jacques.

David Booth wrote:

> 
> Oh, and along with that, the sentence just before Figure 1-1 states:
> 
> "WSDL further considers that an interface represents (some of) the
> behavior of a resource on the World-Wide-Web."
> 
> I don't think that's accurate.  For one thing, it is the _service_
> that is related to the resource (as explained in my previous message
> below), and for another thing, that resource might not be on the Web.
> 
> 
> I suggest rewording this as:
> 
> "WSDL further considers that a service may actually be providing 
> access to some other resource[1], where this other resource may be a
> physical object or something abstract or conceptual.  Thus, multiple
> services may be related by accessing the same resource.
> 
> Comments?

Received on Friday, 20 June 2003 12:17:59 UTC