W3C

WS Architecture Telcon
8 May 2003

Attendees

Present: Abbie Barbir, Dave Orchard, Daniel Austin, Dave Hollander, Doug Bunting, frank mccabe, geoff arnold, Hao He, Katia Sycara, Heather Kreger, Martin Chapman, Mike Champion, Paul Denning, Roger Cutler, Tom Carroll, Ugo Corda, Hugo Haas

Regrets:

Chair: Mike Champion

Scribe: Paul Denning

Contents


 
------------------------------------------------------ 
1. [15:35] Confirm scribe.  The current  list is at 
http://www.w3.org/2002/ws/arch/2/04/scribes-list.html
 
Paul Denning
------------------------------------------------------ 
2. Administrative matters

2a. [15:37] Approve last week's minutes
http://www.w3.org/2002/ws/arch/3/04/2003-05-01-ws-arch.htm
(not yet up, try early Thursday)
::::::::::::::::::::::::::::::
Hugo to enable access.

2b. F2F Logistics

There will be a general strike in France on Tuesday, the
day on which several of us will arrive.  It appears that the
high-speed trains will be affected.  How about a quick exchange of 
schedules and meeting points for carpooling?  Advice from Hugo on 
what to expect and how to cope?
::::::::::::::::::::::::::::::
20030508T1538

Discussion of possible train strike affecting F2F travel in France

Carpooling to Rennes?
MC car.  Leaving late morning.  Interested riders: Abi, Doug Bunting (arriving around 2)
Should not plan to reply on subway; it will not work.

CDG meeting place?  Meet at train station?

Air France may also strike, but Hugo has not heard of Air France strike

Domestic flights usually leave from other airport (not CDG)
There is a flight to Rennes from CDG, approx US$500

CDG north of Paris
Other airport south; hard to get there if train strike.

DO: WSD agenda Nov F2F US west coast or Australia?  Discuss at F2F

Map of TGV?

RER and TGV are different stations.
TGV next to 2S?
Take bus between terminals

Update:  strike to include (possibly) Air France and air traffic controllers

Contact travel agent

2c. XKMS Reviewer
We were not specifically requested to submit feedback for the 
XML Key Management Specification, but is should be on our
radar and we should make sure we don't have any complaints.
http://lists.w3.org/Archives/Member/chairs/2003AprJun/0034.html

The document address is:
  http://www.w3.org/TR/2003/WD-xkms2-20030418/Overview.html
  http://www.w3.org/TR/2003/WD-xkms2-bindings-20030418/Overview.html
  
The Last Call review period will end on 23 May, 2003. Will anyone
volunteer to review these documents and advise the WSA WG of 
any issues that we might want to bring to their attention?
::::::::::::::::::::::::::::::
T1552.
Anything questionable?
Anything to harvest?

SOAP 1.2 and 1.1 bindings.

No mandate from WSAWG to review, no volunteers, but would be nice.
Read during travel to F2F?

------------------------------------------------------
3. [15:50] Action Item Review 
::::::::::::::::::::::::::::::
see http://www.w3.org/2003/05/01-ws-arch-irc

ACTION: Mike Champion to schedule overlapping sessions on WSDL terminology and XML subsetting issues [1] 
[close]

ACTION: Martin to UMLize spaghetti [2] 
[ongoing]

ACTION: Geoff will tweak text of asynch and synch MEPs along the lines of this discussion and resubmit [3] 
[close]
[new ACTION] (Geoff and Frank):  check WD text for most recent
ed note now to include results of async discussion
Mark Jones layered conversation
refactoring, not straight cut and paste
refactor for next WD

ACTION: Glossary editor will incorporate Geoff's MEP definitions [4] 
[ongoing]

ACTION: Geoff will add language in response to Ugo's point cautioning not to use this informal definition in formal specs [5] 

work on Conversations may help resolve async discussion
not a rigorous definintion.
read the ed notes!

ACTION: Take this discussion of MEP taxonomy to mailing list, Suresh will respond on email thread [6] 
[closed] http://lists.w3.org/Archives/Public/www-ws-arch/2003May/0030.html

ACTION: Glossary and document editors need to make sure that the synch/asynch text is in synch [7] 
[ongoing]

ACTION: Group should review frank's latest draft in next few days and record "showstoper" problems [8] 
[ongoing]
No show stoppers reported.  Still time to add ed notes.  Hugo to prepare submission Monday morning (France time) and submit later in the day.

ACTION: MTF people should look carefully at Frank's translation of their text into the concepts relatioships framework [9] 
[ongoing]

ACTION: People should look carefully at message reliability section [10] 
[ongoing]

ACTION: The security and privacy sections need careful attention by group and editors [11]
[ongoing]
------------------------------------------------------
4.[16:00] Publishing the documents 

We will be publishing the following as Working Drafts early next week.
Please review and suggest "warning" flags for bits that are controversial or
rough.
There is still time to make minor changes if specific text is suggested.

http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/glossary/wsa-glossary.html
http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/scenarios/ws-arch-scenarios
.html
http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/wsa/wd-wsa-arch-review2.html

Some issues:

How about the Constraints section ... it's pretty sparse.

??? Raise some "should we do X before publishing" questions!
::::::::::::::::::::::::::::::

Request for pub on Monday 12 May.
Time for minor editorial changes; suggest text
can add ed notes if you are uncomfortable with something

Nothing on security in doc

DO has problem with Abi draft

Frank to add placeholder/ed note with link to Abi's text

Igor would like to see his contributions discussed.

Need to have security discussion at F2F

Gerald - 

Problems with Abi's security write up not  on list, but discussed in telecons

Will not resolve before WD is published.

Daniel to do make updates in time for Hugo to prepare submission.


Usage scenario updated, not major.  Not ready for publication now, or publish what we have.
No less rough than glossary
Any dissent on publication?  no dissent
Hugo will take care of it.
DO work with travel scenario?  Static version.  Also REST and non-REST scenario in the works.

MC: two sections Mike submitted in last week (stack diagram wording, web service 

What do we call non-REST web services?

REST = generic or uniform interface
Opposite is application-specific, known as (API-SOA)
uniform operations (CRUD or PUT POST DEL GET) well-understood (no WSDL or other formal description)
API_SOA needs WSDL (or other 

Frank, others do not like "API"

Can't agree on def for web service, therefore need to add precision, but trip over the labels.
Long-winded, tedious labels

Use "name spaces" to distinguish.

Martin: Scope still up in the air (ebXML, REST, 

See David Booth suggestion a few weeks ago - just say what 
Chris objected to making it too broad.

How does what we are doing relate to the web and REST.

Paul: WS-I does "profiles".  W3C/WSAWG should not be preceived as doing WS-I-"like" profiles, 

Not "Profiles". Rather different "styles".

RESTful SOA means verbs are GET/PUT/...

properties

venn diagram.

Need discussion diagram.

Martin to send an email with his thoughts.

W3C or ebXML style.

Web service with WSDL is not second class citizen (not lesser than restful SOA).

REST versus Application-Specific
REST conceptually bound to HTTP, WSAWG charter not limited to HTTP.

REST says nothing about payload (being XML)

# verbs

Frank: Agents, messaging level of discussion is different than REST/verbs level of discussion

REST -> HTTP -> synchronous

HTTP is an incarnation of REST

Roger: sync vs async is a big distinction between REST and other styles.

Frank:  taking it too far.

DO: term Generic method SOA, 

Paul:  open vs limited?

DO "extensible"

"using" relationship

uniform vs open

add ed note to warn that terminology is open, but concept is more solid

remove "API"

Use "uniform",  uniform-interface services (UIS)

Use "open-interface services" (OIS) for non-Restful

UIS is an instance of OIS with restrictions added.  "uses" relationship.

----------------------------------------------------
5. [16:15] F2F Agenda review
http://lists.w3.org/Archives/Member/w3c-ws-arch/2003May/0024.html

Please suggest any changes.  

The basic "theme" of the meeting is to identify what we must do to get
from the new WDs to Last Call.  This involves both determining what we
need to do that we haven't really discussed yet, and to assess the state
of the WSA documents. As a firt step, we will do "triage" on the documents:

Identify those parts that are basically OK and need simple wordsmithing; 
identify those that need a  lot of work, and identify those on which 
there is little hope of consensus.  We will then give priority to the
cases for which there is hope of consensus but work to do; the wordsmithing
can be done as a background task by editors and "issue owners" and the
meanest, nastiest trout can be addressed when we have time and (hopefully)
some momentum toward a final document.  

So, the discussions will be very focused on the content of the document,
not at the level of wordsmithing, but at the level of concepts and 
relationships.
::::::::::::::::::::::::::::::
[T1652] 

(1) What is in and out of scope?

SOAP - definitely in.
ebXML - ?

(names we use will give impression of scope)

"scope" is different than "meaning"

(2) Triage 
Identify or classify document sections that are okay versus in need of work.
Conceptually sound versus word-smithing.

What things need a lot of work to reach concensus?

(3)  Work on the hard, but manageable problems.
F2F best to address hard conceptual stuff.  Word-smithing better for email/telecon.

Topics:
Security (need experts)  Hal Lockhart BEA comments on Abi draft to be sent.

MTF regrets for F2F.

4pm in France

Fri afternoon trains

Remote participation:

Canon to set up IRC for minutes, and catch up of progress

There is a 6 hour time delta from US East Coast

Telephone Bridge in France.

------------------------------------------------------
6. [16:45 ] Other pending issues (if time available)

State of the Section 1.6.2/3 text ... open issues?
What do we call a non-REST Web service?
::::::::::::::::::::::::::::::
Adjourn: 1705.

<frankmccabe> Comment: endorsing REST has significant issues wrt HTTP and protocol independence

<DaveH> p+
... suggest - sync/async may be easier to understand if you say "stateful"

<frankmccabe> generic soa, rest soa

Summary of Action Items


David Booth
dbooth@w3.org
$Date: 2003/06/12 14:22:25 $