Re: Draft of the Web Services Glossary

I have integrated the changes below in:

  http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/glossary/wsa-glossary.html?rev=1.41&content-type=text/html&only_with_tag=edcopy_2003050601

* Cutler, Roger (RogerCutler) <RogerCutler@chevrontexaco.com> [2003-04-30 14:03-0500]
> Here are some comments.  In general, portions of this are MUCH improved
> -- but some of the sections are weird, to put it mildly.  Would it be
> possible to flag the sections that are included for some "to-be-done"
> reason and distinguish them from sections that have been reasonably
> worked over?

I guess that you are referring to the core concepts section[1] and the
choreagraphy section[2].

I have marked them both as under heavy rework. Does that address your
concern?

> Actor - As I understand it a legal entity can own agents or Web services
> separately.  That is, ownership of a Web service does not always come
> through ownership of the agent, and in fact the two may have different
> owners.  I think.  Suggest: "... That may be the owner of agents and/or
> Web services".

Hmmm... actually, worded as such, actor is a synonym of legal
entity[3].

I don't think that the definition of actor needs to be changed, but
rather actor should be replaced by legal entity. I have added such a
comment in the editorial note about that for legal entity.

> Discovery - Is this compatible with David Booth's diagrams and analysis?
> In particular, I am concerned that some scenarios do not involve machine
> processable descriptions but instead what he calls "semantics".
> Suggest: "The act of locating a description ..."  At the very least I
> would like to see this flagged as not representing consensus.

This is a definition that David came up with, so I trust it reflects
its work. It is actually used in the architecture document.

Does that address your concern or do you want me to add an editors'
note?

> Message - I question the use of the word "client", which has an
> implication to me of a client-server model.  The definition of "client",
> in fact, simply refers to "requestor".  Suggest:  Eliminate client in
> favor of requestor, see if one can eliminate client altogether.  Also,
> the English is fractured somehow unless there is some punctuation after
> the word client that I cannot see in this font. 

Agreed. Client has disappeared and been replaced by requester
everywhere.

> Safe - Is this consistent with other sources, notable Web architecture?
> If so, should reference.

This definition is adapted from RFC2616, so I trust that it should be
consistent with the Web architecture. I have added an RFC2616
reference.

Same comment for idempotent.

> Web site - Is this accurate?  Does it come from somewhere?  It seems to
> me that Web sites can include things that are not pages, like
> executables and Web services.

As noted, this definition comes from Web Characterization Terminology
& Definitions Sheet[4].

Web site is only used by browser, which isn't used anywhere else. The
term appears in the architecture document, but maybe the term is
self-explanatory.

If there is much discomfort, I can remove those two terms altogether.

> Legal Entity - Suggest: "... Or of Web services themselves".

This is extracted from the architecture document. I'd rather leave it
alone for now unless you can't live with it.

> Manageable element - Suggest eliminate or improve.  This definition adds
> nothing that is not evident from the term itself.

I added an editors' note to that effect. Again, it is extracted from
the architecture document. I have updated the definition which has
changed after Frank's edits.

> Manageability Interface - This seems too general to me.  I think it
> covers way too much, as I understand it.  Suggest eliminate or rework.
>
> Message Description Language - Circular.  Eliminate.
> 
> Message identifier, recipient, etc -- Ditto.

Same comment.

> Declarative and Procedural - WHAT????  I give up -- this is getting too
> weird.

This section is, I think, clearly marked as in progress. I agree that
those two definitions are less ready than the others. I have commented
them out.

* Jon Dart <jdart@tibco.com> [2003-04-30 12:13-0700]
> I trust the "reliable messaging" definition is still a work in progress.
> 
> As it stands, the text says that RM implies both confirmation of receipt 
> and once-and-only-once delivery. In fact, real reliable message systems 
> may offer either of these capabilities, or both, and in addition other 
> capabilities, as part of a spectrum of quality of service options.
* Cutler, Roger (RogerCutler) <RogerCutler@chevrontexaco.com> [2003-05-01 11:34-0500]
> In general this looks pretty good to me, although some wordsmithing and
> smoothing is obviously still necessary.  I agree with both of Jon's
> comments below.
> 
> Hugo -- you should note that the glossary definition of RM is now out of
> whack with the document.

Done.

Regards,

Hugo

  1. http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/glossary/wsa-glossary.html?rev=1.40&content-type=text/html&only_with_tag=edcopy_20030430#coreconcepts
  2. http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/glossary/wsa-glossary.html?rev=1.40&content-type=text/html&only_with_tag=edcopy_20030430#choreographydefs
  3. http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/wsa/wd-wsa-arch-review2.html?rev=1.13&content-type=text/html#legal_entity
  4. http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/glossary/wsa-glossary.html?rev=1.40&content-type=text/html&only_with_tag=edcopy_20030430#WCTDS
-- 
Hugo Haas - W3C
mailto:hugo@w3.org - http://www.w3.org/People/Hugo/

Received on Tuesday, 6 May 2003 05:35:21 UTC