W3C home > Mailing lists > Public > www-ws-arch@w3.org > March 2003

RE: Questions prompted by the publication of WS-ReliableMessaging

From: Ugo Corda <UCorda@SeeBeyond.com>
Date: Mon, 17 Mar 2003 13:18:16 -0800
Message-ID: <EDDE2977F3D216428E903370E3EBDDC9081126@MAIL01.stc.com>
To: "David Orchard" <dorchard@bea.com>, <www-ws-arch@w3.org>
I am personally not concerned at all whether reliability is handled by OASIS instead of W3C. What really worries me in general is that duplicate work is being done in competing groups. The publication of WS-ReliableMessaging without any clear indication of plans to converge with the OASIS work (publication made, by the way, by companies that, as far as I know, are not part of the OASIS reliability TC so far) definitely raised my worries quite a bit. 
 
If you say that there is a good possibility that WS-ReliableMessaging will be submitted to the OASIS reliability TC, that makes me feel better. But I'll be looking for concrete actions and not just words. 
 
By the way, it puzzles me that this submission has not yet occurred, if the authors indeed contemplate that possibility. The OASIS TC just started its work, and this would be a great time for submission: members could start right away to look at both specs and get the best of both, instead of starting work in one direction and having to change course later.
 
Ugo

-----Original Message-----
From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]On Behalf Of David Orchard
Sent: Monday, March 17, 2003 12:40 PM
To: www-ws-arch@w3.org
Subject: RE: Questions prompted by the publication of WS-ReliableMessaging


I think that companies working on WS-ReliableMessaging bodes well for the widespread adoption of an sigle, open, RF, interoperable reliability specification that architects well with other Web service specifications.  
 
What I'm really surprised by is the pushback that people are giving on the possibility that ws-rm won't automatically go to the oasis ws-reliability TC.   One of the biggest differences between the W3C and OASIS is that the W3C has a community review process for new working groups.  The goal at the W3C being to have a charter for a working group that all the W3C Members can live with and is technically consistent with other works in progress.  Indeed, the work that the W3C Team and member companies do to come up with such a charter and then the 4 week AC review period are the biggest contributors to the length of time to form a Working Group, IMO.  OASIS has no such intent or process.   
 
If you like forming WG's quickly, you can't ALSO want there to be consensus in the industry around a single Working Group and it's charter.  Time to market or consensus, pick one.  
 
Continuing the community review of charters a little further, TC's can arbitrarily change their charter without a community review.  The IP issues around changes in charters alone gives me the willies, like how the heck do you do early IP disclosure if the charter can change???
 

Further, individual members of OASIS don't treat TC formation as a "first to pole wins" wrt charters.  OASIS provides a place for like minded individuals to do work.  It has no structure for insisting on consistency or relationship between TCs or other groups.  Though such is certainly permitted, pending whether the individuals choose to.  
 
Surely these differences in processes and their plus and minues have been obvious for years now and people shouldn't be surprised by what can happen, or that it automatically means something *bad* for Web services.  If people have reservations about the OASIS process, then talking about it on w3c ws-arch seems like it would need to get moved off of the w3c pretty fast.  I'm ok with talking about the impact of various processes on web services architecture for a while, but it's probably something that needs to be wrapped up quickly.  Finally, people seem to like the OASIS process as lots of work is being done there.  
 
Cheers,
Dave

-----Original Message-----
From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]
Sent: Saturday, March 15, 2003 2:34 PM
To: www-ws-arch-request@w3.org; www-ws-arch-request@w3.org; 'Ugo Corda'; www-ws-arch@w3.org
Subject: RE: Questions prompted by the publication of WS-ReliableMessaging


There are differences. The primary difference that I see (after a quick glance) is that WS-ReliableMessaging relies on/works with WS-Policy and the other GXA specifications. WS-Reliability is a standalone SOAP extension. Also, WS-ReliableMessaging has defined Addressability separately (which decouples asynchronicity from reliability). I think that WS-ReliableMessaging and WS-Addressability are better, more flexible, more thorough, more comprehensive specifications. Even so, the specs address exactly the same problem space. 
 
The primary difference is political. The authors of WS-ReliableMessaging have not signed up to participate in the WS-RM TC. I'm not sure that it's a given that the authors will submit it to the WS-RM TC. I'd say that it bodes badly for the standardization effort. 
 
Anne

-----Original Message-----
From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]
Sent: Saturday, March 15, 2003 4:05 PM
To: www-ws-arch-request@w3.org; 'Ugo Corda'; www-ws-arch@w3.org
Subject: RE: Questions prompted by the publication of WS-ReliableMessaging


I agree with Ugo. Reading through the abstract it's obvious that we have two specifications that solve the same problem. Is there a value in that?
 
I actually dug deeper into the specs and I can tell that there are some differences. But most of us don't have the time to compare green apples to red apples. It would have been much easier if someone could present a list of the difference. If WS-ReliableMessaging does something better than WS-RM then clearly it could be summarized in two pages and presented to the WS community so we can judge.
 
Maybe they are so different that we need to have both. I don't see that, but a more educated explanation would help. Maybe the changes are minor, in which case such a comparison could help the OASIS TC in addressing the problem of reliable messaging in a much better way.
 
Am I the only one interested in seeing such a comparison?
 
arkin

-----Original Message-----
From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]
Sent: Saturday, March 15, 2003 12:46 PM
To: 'Ugo Corda'; www-ws-arch@w3.org
Subject: RE: Questions prompted by the publication of WS-ReliableMessaging


I suggest you need to read the specs slower rather than quicker :-)

-----Original Message-----
From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]On Behalf Of Ugo Corda
Sent: Saturday, March 15, 2003 12:44 PM
To: www-ws-arch@w3.org
Subject: Questions prompted by the publication of WS-ReliableMessaging



Probably most people in the group have had a chance by now to see this week's announcement of the publication of WS-ReliableMessaging (see [1]).

After a quick reading of the spec, I have to say that I don't see any major architectural/technical differences compared to the OASIS WS-ReliableMessaging TC activity and its input document WS-Reliability (or at least differences big enough to justify going a completely separate way). 

I really hope that some WSA members whose companies published the new reliability spec can help me clarify the previous point and provide some architectural/technical rationale for the separate publication.

Thank you, 
Ugo 

P.S. No need to answer if the rationale for publication is a political one (I can figure that out by myself ...). 

[1] http://xml.coverpages.org/ni2003-03-13-a.html 
Received on Monday, 17 March 2003 16:18:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:16 GMT