IRC log of ws-arch on 2002-12-19

Timestamps are in UTC.

20:34:55 [RRSAgent]
RRSAgent has joined #ws-arch
20:34:59 [ugo]
ugo has joined #ws-arch
20:35:18 [zulah]
zulah has joined #ws-arch
20:35:25 [Kreger]
Kreger has joined #ws-arch
20:35:44 [Roger]
zakim, who is here?
20:35:45 [Zakim]
sorry, Roger, I don't know what conference this is
20:35:46 [Zakim]
On IRC I see Kreger, zulah, ugo, RRSAgent, Zakim, JimD, Roger, MartinC, mchampion, chrisf, Mark_J
20:36:13 [prasad]
prasad has joined #ws-arch
20:36:24 [chrisf]
zakim, this is arch
20:36:25 [Zakim]
ok, chrisf
20:36:32 [Roger]
zakim, this is ws
20:36:34 [Zakim]
this was already WS_ArchWG()3:30PM
20:36:34 [Zakim]
ok, Roger
20:36:34 [Zakim]
+M.Mahan
20:36:45 [Zakim]
+ +1.650.849.aabb
20:36:47 [Roger]
zakim, who is here?
20:36:49 [Zakim]
On the phone I see Prasad_Yendluri, ??P9, Mike_Champion, ??P11, DOrchard, ??P14, ??P15, Hugo, +1.650.875.aaaa, Colleen_Evans, Mark_A_Jones, Eric, ??P20, ??P21, ??P22, ??P24,
20:36:50 [Zakim]
... Chris_Ferris, M.Mahan, +1.650.849.aabb
20:36:52 [Zakim]
On IRC I see prasad, Kreger, zulah, ugo, RRSAgent, Zakim, JimD, Roger, MartinC, mchampion, chrisf, Mark_J
20:37:12 [Zakim]
-??P24
20:37:13 [Zakim]
-??P11
20:37:23 [Zakim]
+??P27
20:37:33 [Zakim]
+??P11
20:37:34 [chrisf]
zakim, aabb is probably sinisa
20:37:35 [Zakim]
+Sinisa?; got it
20:37:45 [Roger]
zakim ??p11 is Roger
20:37:57 [Roger]
Zakim, ??p11 is Roger
20:37:58 [Zakim]
+Roger; got it
20:38:52 [JimD]
having some cell phone troubles; will reconnect in 5
20:40:49 [Roger]
Roger scribing.
20:41:22 [Roger]
A bunch of minutes have been posted. Hugo wants more time to review.
20:41:49 [Roger]
Clean up response to XMLP attachment action item?
20:41:52 [Roger]
Not done yet.
20:42:12 [Roger]
WSDL requirements draft -- any architectural issues noted?
20:42:20 [Zakim]
+??P24
20:42:37 [JimD]
??P24 is JimD
20:42:51 [Roger]
Chris F volunteers to do over holidays. WSDL wants response by 12/31, but they can probably get whatw they get.
20:43:07 [chrisf]
ACTION: Chris to review WSD Requirements WD and summarize any potential arch issues
20:43:27 [Roger]
Note sent to Mark Baker saying will consider his issue, we are not closing it.
20:43:29 [JimD]
zakim, ??P24 is JimD
20:43:31 [Zakim]
+JimD; got it
20:43:54 [Roger]
Mike C has volunteered to harvest RM threads for document.
20:44:06 [chrisf]
zakim, who is here?
20:44:07 [Zakim]
On the phone I see Prasad_Yendluri, ??P9, Mike_Champion, DOrchard, ??P14, ??P15, Hugo, +1.650.875.aaaa, Colleen_Evans, Mark_A_Jones, Eric, ??P20, ??P21, ??P22, Chris_Ferris,
20:44:09 [Zakim]
... M.Mahan, Sinisa?, ??P27, Roger, JimD
20:44:10 [Zakim]
On IRC I see prasad, Kreger, zulah, ugo, RRSAgent, Zakim, JimD, Roger, MartinC, mchampion, chrisf, Mark_J
20:44:48 [Roger]
Hugo has on to-do list to take care of capitalization of wEB sERVICES in glossary, etc.
20:44:58 [Roger]
He's going to put everything lower case for the moment.
20:45:31 [Roger]
It was discussed last week. WSDL wants some sort of input pretty soon.
20:45:56 [Roger]
We decided "Web services" last week, we think.
20:46:19 [Roger]
ACTION: Hugo to clean up capitalization of web services in glossary.
20:46:38 [Roger]
MTF report?
20:46:39 [Zakim]
+??P28
20:47:03 [Roger]
Heather - discussions of requirements for information on WS components ongoing.
20:47:06 [Roger]
Or something like that.
20:47:10 [chrisf]
zakim, ??p28 is igor
20:47:11 [Zakim]
+Igor; got it
20:47:34 [Shishir]
Shishir has joined #ws-arch
20:47:51 [Roger]
Tom has published the issues list, the URL is in the agenda.
20:48:13 [Roger]
We will now plan on structuring time in the telcons for processing issue list.
20:49:05 [Roger]
Mark J - What are these issues against? General? Arch doc? Requirements doc?
20:50:02 [Roger]
Chris - Issues list should indicate which doc, if any, it applies to. This appears to be the "Req" column, which seems to have such references in it.
20:50:20 [Roger]
Action: Tom clarify what documents are referred to.
20:50:40 [Roger]
ACTION: Tom clarify what documents are referred to.
20:50:50 [Roger]
Who is Tom?
20:52:05 [Zakim]
-M.Mahan
20:52:06 [chrisf]
ACTION: TomC to clean up issues list so that the Spec column uses consistent terms (e.g. both Arch and Spec are used to refer to the arch WD)
20:52:29 [Roger]
Actions 3 and 4 are the same.
20:52:33 [Zakim]
+M.Mahan
20:53:10 [Roger]
If you find yourself in need of something to do, go through issues and think about draft responses. Start discussion if seems warrented.
20:53:37 [Roger]
Reliability:
20:54:23 [Roger]
Roadmap for major issues and status?
20:54:44 [Roger]
1 - Security. Status: deferred to OASIS for the moment.
20:55:02 [Roger]
IBM et al has released a bunch of security specs.
20:55:17 [Roger]
Should we be looking at these and referencing them?
20:56:06 [Roger]
Seems like a good idea to be familiar with them.
20:57:12 [Roger]
Microsoft and IBM are both buying into and implementing this security roadmap.
20:58:13 [chrisf]
here's the link to the updated roadmap and all of the recently published Policy and Security specs: http://www-106.ibm.com/developerworks/library/ws-secroad/
20:58:21 [Roger]
ACTION: MikeC to summarize RM threads for Arch document.
20:59:25 [Roger]
2 - Choreography. Status - in whatever state it's in, probably waiting for TB-L.
20:59:31 [Roger]
3 - Reliable Messaging
21:00:07 [Roger]
What other things besides RM should we be thinking about besides RM as focii of attention.
21:00:48 [Roger]
Somebody asks whether a whole WG is needed for a simple ack?
21:01:18 [Roger]
Duane Nickol - Requirements of business often both RM component AND a service agreement saying what happen if fails.
21:01:57 [Roger]
If scoping RM, should leave hook in for something to describe those rules of engagement.
21:02:12 [Roger]
Same somebody says it looks like this is overlapping with choreography.
21:03:09 [Roger]
DN - Rules of engagement are next level up. One approach is to keep RM down the stack and push the agreement component events up the stack.
21:03:33 [Mark_J]
q+
21:04:05 [mchampion]
ack mark
21:04:08 [mchampion]
ack hugo
21:04:36 [Roger]
MarkJ - MOM lets client send message and have something higher up stack do best to get delivered.
21:04:38 [MartinC]
q+
21:04:50 [Roger]
Message queuing and so on.
21:05:06 [Roger]
Is that a feature of asynch messaging or RM? Or both?
21:05:45 [Roger]
Somebody says they have tried to make distinction between business level ack and transport level ack.
21:06:02 [Roger]
Transport level is closer to synchronous, pushes info up stack.
21:06:22 [Roger]
Business level acknowledgement is at a higher level, involving agreed choreography.
21:06:27 [Roger]
q+
21:07:09 [Roger]
Software that tries to get delivery without dealing with business level?
21:07:38 [Roger]
ebXML relies on TCP/IP to get simple response code. Higher level tracks through to business level choreography.
21:07:56 [Zakim]
-M.Mahan
21:08:02 [Roger]
Does the business level get involved immediately if synchronous transport fail?
21:08:17 [Roger]
Transport level does not require immediate success.
21:08:34 [chrisf]
can we have one conversation please?
21:08:56 [Roger]
Reliability processes usually concentrate on transport.
21:09:20 [Roger]
I have no earthly idea who is talking.
21:09:30 [Zakim]
+??P5
21:09:59 [Roger]
From a business level, you often need to send a separate business level ack back.
21:10:21 [Roger]
q?
21:10:33 [Zakim]
+M.Mahan
21:11:19 [mchampion]
ack martinc
21:11:44 [Roger]
MartinC - Another thing, related, is other MEP's. Asynch, once only, etc.
21:11:58 [Roger]
One to many, intermediaries, etc.
21:12:06 [Roger]
THis is something else for the roadmap.
21:12:18 [mchampion]
ack roger
21:13:33 [Roger]
Roger - separate business from message ack.
21:14:42 [Roger]
MikeC - couple issues on the table.
21:14:56 [Roger]
Is this the next thing we want to table or do we need to sort out MEP's first.
21:15:30 [Roger]
Is this something where there is a simple, 80-20 -- low hanging fruit -- simple protocol that retries N times?
21:15:48 [Roger]
Is it worthwhile to spawn off a spec in this simple vein?
21:16:16 [Roger]
How do we sort out more complex questions of how to deal with failures - choreography and business level implications.
21:16:39 [Roger]
Certainly would not dump the latter on XMLP, but the simple, low-fruit might be a possibility.
21:17:16 [Roger]
Chris - Thinks the two issues are orthogonal. Msg gets through or doesn't. That's one issue. Implication of whether the msg got through is another issue.
21:17:58 [Roger]
DN - ebXML agrees that failure is failure -- has hook to go to what implication is.
21:18:29 [Roger]
What is the mechanism for going ahead on the simple thing? Full WG seems to be overkill. Will XMLP take it?
21:19:33 [Roger]
MikeC - Has this come up in XMLP meetings?
21:19:50 [Roger]
They are clearly going to work on attachments?
21:20:22 [Roger]
Call yesterday - there's a laundry list that the group could do, but the group does not feel well positioned to determine priorities.
21:20:46 [Roger]
Would like input on prioritization from WS-ARCH or CG.
21:20:55 [Roger]
Also potentially chartering issues.
21:21:09 [Roger]
I think the previous comments were from MarkJ.
21:21:33 [Roger]
Mike - if WS-ARCH said this was an important thing, what would XMLP think?
21:21:54 [Roger]
Answer - Would probably have issues about doing it without going to AC.
21:22:04 [Roger]
Everybody agrees -- would have to go to AC.
21:22:29 [Roger]
q+
21:22:56 [Roger]
Somebody says should do architecture, not RM.
21:23:20 [chrisf]
+1 to that
21:23:26 [Zakim]
-Chris_Ferris
21:23:36 [DaveO]
DaveO has joined #ws-arch
21:23:39 [Roger]
DN - OASIS TC would be favorable to receiving comments from this group. Messaging Group.
21:24:47 [Roger]
The ebXML messaging group may be evolving toward a more generic messaging group.
21:25:26 [Roger]
Coleen Evans - Msging group does not seem to be heading that way yet.
21:26:13 [Roger]
Is that group actually going there? Probably more discussion than actual motion.
21:30:20 [Roger]
Roger - Low hanging fruit should be grabbed and handed off.
21:31:49 [Roger]
MikeC - Concentrate on arch, defer RM until F2F.
21:31:56 [Roger]
q?
21:32:01 [Roger]
q-
21:33:33 [Roger]
Hao - RM experience - Time factor helps to simplify situation dramatically.
21:34:31 [Roger]
Simpler if responsibility for RM stays with client.
21:34:33 [Zakim]
+??P0
21:35:06 [Zakim]
-Prasad_Yendluri
21:35:11 [dougb]
dougb has joined #ws-arch
21:35:30 [Roger]
Applications can simply take care of this without a protocol?
21:35:45 [Roger]
Or is there value in simple re-try protocol?
21:41:44 [Roger]
Hao is saying that sender has the basic responsibility in RM.
21:42:00 [Roger]
Or that there is an RM mechanism that will work where this is true.
21:42:17 [Roger]
We seem to be talking past each other -- seems to be disconnect here.
21:42:26 [Roger]
Need to discuss on mailing group.
21:43:55 [Mark_J]
q+
21:45:15 [Roger]
ebXML spec. Need URL for it.
21:45:35 [mchampion]
ack mark
21:47:01 [Roger]
Levels seem to hurt.
21:47:21 [Roger]
Need to list features that are needed rather than immediately try to categorize them.
21:47:27 [RRSAgent]
See http://www.w3.org/2002/12/19-ws-arch-irc#T21-47-21
21:48:29 [Roger]
What does an ack mean? Different aspects there as well.
21:48:47 [Roger]
ebXML: couples meaning of ack with behavior of application.
21:49:03 [Roger]
Specifies that the message handler will order messages.
21:49:23 [Roger]
ebXML is more than messaging but also has components of interface to apps.
21:49:37 [Roger]
Could talk about communication protocol on its own.
21:51:12 [Roger]
Here is the URL for the ebXML Messaging Spec V2: http://www.oasis-open.org/committees/ebxml-msg/documents/ebMS_v2_0.pdf
21:52:02 [Roger]
Tuning of protocol for performance can be separated out, too.
21:52:27 [Roger]
Three things; 1 - Bits on the wire exchanged; 2 - TUning of exchange; 3 - QOS offered on either side.
21:52:31 [dougb]
The ebXML Messaging 2.0 specification is also available at http://www.ebxml.org/specs/ebMS2.pdf
21:53:12 [Zakim]
-M.Mahan
21:54:20 [Roger]
Look around pg 35 of the ebXML spec.
21:55:28 [Roger]
MikeC has the action item to harvest the mailing list and references to take a first cut at architecture straw man.
21:55:36 [Roger]
For RM.
21:55:56 [Roger]
New Topic.
21:56:10 [Roger]
Comment from P3P WG saying that they don't get no respect.
21:56:30 [Roger]
Is this an official WG communication?
21:56:44 [Roger]
Hugo - This issue has some history.
21:57:17 [Roger]
P3P had some requests from XMLP. Wanted to be able to associate P3P policy with SOAP message.
21:57:34 [Roger]
XMLP said that they weren't going to provide the explicit connection.
21:57:46 [Roger]
Passed along to Arch group.
21:58:06 [Mark_J]
q+
21:58:30 [Roger]
q+
21:59:08 [Roger]
Need to have on issues list.
21:59:19 [Roger]
MarkJ - Concerns are deeper and have arch implications.
21:59:39 [Roger]
P3P policies refer to URI's. This talks about a particular operation or service.
21:59:50 [Roger]
No canonical way to describe implications of service.
22:00:06 [Roger]
How do you state that a SOAP message will satisfy some privacy requirement.
22:00:23 [Zakim]
-??P9
22:01:10 [Roger]
Might get some folks together to consider issue?
22:01:24 [Roger]
Get P3P people to make presentation at F2F about their concerns.
22:01:32 [mchampion]
ack mark
22:02:12 [mchampion]
ack roger
22:04:28 [Zakim]
-??P27
22:04:48 [Roger]
Mike - schedule face time at Plenary?
22:04:57 [Roger]
Hugo - Seems reasonable.
22:05:40 [Zakim]
-DOrchard
22:06:31 [Roger]
ACTION: Hugo coordinate with P3P group and schedule communication.
22:06:36 [Zakim]
-JimD
22:06:39 [mchampion]
ACTION: Hugo will contact P3P people to suggest meeting at Plenary in march
22:07:14 [Zakim]
-Colleen_Evans
22:07:14 [Zakim]
-??P21
22:07:15 [Zakim]
-Igor
22:07:16 [Zakim]
-Mark_A_Jones
22:07:18 [Zakim]
-??P5
22:07:19 [Zakim]
-??P22
22:07:20 [Zakim]
-Sinisa?
22:07:21 [Zakim]
-??P15
22:07:21 [Zakim]
-??P14
22:07:24 [Zakim]
- +1.650.875.aaaa
22:07:25 [Zakim]
-??P20
22:07:34 [Zakim]
-Hugo
22:07:35 [Zakim]
-Eric
22:08:42 [Zakim]
-Roger
22:10:41 [Zakim]
-??P0
22:10:42 [Zakim]
-Mike_Champion
22:10:44 [Zakim]
WS_ArchWG()3:30PM has ended
23:03:48 [ugo]
ugo has left #ws-arch