Re: Snapshot of Web Services Glossary

Hi Jean-Jacques.

Thank you for your comments.

* Jean-Jacques Moreau <jean-jacques.moreau@crf.canon.fr> [2003-02-14 11:44+0100]
> "binding"
> =========
> I think you need to differentiate between:
> 
> a) the binding of SOAP to HTTP, i.e. how one uses HTTP for 
> transferring a SOAP envelope over the wire (see SOAP Part 2, 
> section 7, SOAP HTTP Binding [1]).
> 
> b) the WSDL SOAP binding, i.e. how an abstract WSDL message is 
> mapped to SOAP (see WSDL Part 2, section 2, SOAP Binding [2]).
> 
> Currently, b) is not covered by the glossary. "binding, 1." 
> describes an abstract (WSDL) binding. I suggest you add "binding, 
> 1'." to describe a concrete (WSDL) binding.

I couldn't find a good way to define a concrete WSDL binding. I looked
at the WSDWG documents and couldn't find a definition either. Is there
one somewhere?

I ended up putting:

  The mapping on an abstract description of a message to a concrete
  protocol.

But I am happy to get something better.

> "endpoint"
> ==========
> The word "binding" is ambiguous. I suggest you refer specifically 
> to a "concrete binding" (definition 1'. above).

I added an editor's note but haven't made the change yet because I am
not totally satisfied by the term "concrete binding".

Again, this is a WSDWG definition. Have these definitions been
updated?

> "safe" and/or "idempotent"
> ==========================
> I suggest you point to add a link to SOAP 1.2 Part 2, section 
> 4.1.2 Distinguishing Resource Retrievals from other RPCs [3].

We definitely need to talk about this, but my feeling is that this
discussion, and reference, should be in the architecture document, not
in the glossary.

> "state"
> =======
> Shouldn't there be a definition for a state machine, as used, for 
> example, to describe SOAP bindings?

I agree that there should actually be a reference somehow. The MTF has
come up with a state machine for the life cycle of a Web service too,
so I will work on that while working on the management definitions.

In the meantime, I have added an editor's note.

> "intermediary"
> ==============
> Any reason not to refer to the SOAP definition?
> 
> "node"
> ======
> Ibid.

I agree with MarkB that those terms need to be defined in a general
way. I have added a reference to the SOAP definition for intermediary.
It was missing.

> "requester"
> ===========
> Ibid. Why use a term different than "sender"? The same comment 
> probably applies to "provider".

Hmmm... sender is really at the message level whereas requester is at
a more conceptual level.

> "role"
> ======
> Ibid.

Please see David B's definition.

> References
> ==========
> The bibref entry for WSD Reqs points to SOAP 1.2 Part 1 (and 
> actually an old version).

Fixed.

Regards,

Hugo

  4. http://www.w3.org/TR/wsdl12-bindings/
-- 
Hugo Haas - W3C
mailto:hugo@w3.org - http://www.w3.org/People/Hugo/

Received on Tuesday, 11 March 2003 15:48:48 UTC