RE: [wsbpel-reqts] notes from our call today

[Diane Jordan <drj@us.ibm.com>]

Peter,
Thanks - sorry you missed the call.
  Most of the focus was on how to organize next week's agenda rather
than the categorization (although no one howled in outrage at my effort).
 In any case,  I don't want to formalize the categories at this
point - there are lots of ways to slice the issues and unless the submitter/champion
wants to adopt one, I don't think the TC should impose one.  Meanwhile,
we can use this informally as one source of input to help plan agendas.
   

We had a lot of discussion about various
ways to run the f2f meeting next week - we didn't really have consensus
on one approach.    For the sake of the folks who were not on
the call (and the memory of those who were ;-) ) here are some notes: 

After administrative start up stuff
and agenda discussion, we would start the meeting with a discussion of
a work plan for the TC (rather than have this at the end as I had first
proposed).   I'm planning to suggest we start iterating on updates
to the spec with the target to have a first pass for review before the
next f2f.  There would be some additional (tbd) number of iterations
before we would have an official draft.  I've yet to review this with
the spec editing team, so it could change.  This might imply that
we need to decide which issues should be included in the first update and
it was suggested we could use the issue discussion on Thurs. to determine
this.   To make room for this, the use case discussion moves to Friday
am.   As in my original proposal, we would then have a discussion
of the goals document and any related questions folks have for Satish Thatte
and Dieter Roller (Frank Leymann will not be present).   Friday's
would cover use cases, liaison reports, spec editing process, and wrap
up.  

Thurs would be dedicated to issue discussions.
 There are a number of options for organizing this day (some we discussed
today and some that were discussed in previous meetings).  We didn't
all agree on all of these: 
1.  "Chunk" the day into
blocks of time allocated for the groupings in the note below.  Since
there are more groups than we would have time to reasonable discuss during
a day we would either choose a subset of the groups or combine some of
them.  I think we all agreed on this but there are a couple proposals
that effect how many we could schedule and how we would use the chunks.
 

2 and 3 are alternate objectives for
the discussions/ways to decide on content:
2.  Spend the discussion time for
a group working on selecting those issues that should be included in a
first pass update of the spec.  This would allow us to finish the
f2f with decisions on what issues need to be addressed immediately, but
discussion of the resolutions would happen after the f2f.   
3.  Pick several important issues
in each category and discuss possible resolutions for them.  This
uses the time for in depth technical exchange but we might have difficulty
deciding which issues are "important" (or more controversially,
describing some set of issues as unimportant).  I have a feeling that
in each group there are only a few issues where folks are really ready
to delve deeply anyway regardless of the importance, so polling champions
and the TC on readiness may be a more neutral selection process. However,
I've already asked for suggestions and haven't gotten much.

4, 5 and 6 are alternative ways to allocate
time:  
 4. Break the group into subgroups
and run parallel discussions on the same topics for 2 hrs., then reconvene
the full group to review and consolidate results for an hour.  This
was discussed in the context of 2 above but could be applied to 3 also.
  The smaller groups might encourage more participation and interactive
discussions.  This limits us to 2 possibly 3 discussion blocks which
would mean the topic groups to be discussed might be somewhat large, or
we wouldn't get to many.  
5. Break the agenda into parallel threads
with different topics running for 1-2 hours and then converging into a
common session for an hour where the different topic groups would summarize
their discussions.   This is more applicable if the content is organized
around 3 above rather than 2.  This limits us to 2 or 3 discussion
blocks, but two (or possibly more) topics could be covered at a time allowing
us to cover more topics.  I've not seen suggestions for which topics
can run in parallel and expect that many folks would want to participate
in conflicting topics.  
6. Run a sequential agenda with the
full group participating in common discussions.  3-5 topics could
probably be scheduled in 1-2 hour blocks (depending on the content).  This
avoids scheduling conflicts, maintains a common base of thinking and doesn't
require time to be spent resyncing the group.   
 
We're open for input and I'm thinking
of sending another note to the TC.   John and I will have to make
a call in order to get the proposed agenda out by Friday but of course
the TC will get to accept or revise it on the first day.    

Let us know what you think.  
 
Regards, Diane
IBM  Dynamic e-business Technologies
drj@us.ibm.com
(919)254-7221 or 8-444-7221, Mobile: 919-624-5123

"Peter Furniss" <peter@furniss.co.uk> 09/10/2003 01:05 PM

To:        Diane Jordan/Raleigh/IBM@IBMUS, "'bpel rqmts'" <wsbpel-reqts@lists.oasis-open.org>         cc:                 Subject:        RE: [wsbpel-reqts] some ideas for our call today

Diane,
 
I'm very sorry, I clean forgot
about the call (I was being useful in other directions).  And I've
also not had time to do the kind correlation survey you've done here, which
looks good.
 
If this categorisation was roughly
what came out of the meeting, I'll add some linking and fields on this
basis to the list.
 
Apologies for my amnesia
 
Peter
 
 
-----Original Message-----
From: Diane Jordan [mailto:drj@us.ibm.com] 
Sent: 10 September 2003 15:26
To: bpel rqmts
Subject: [wsbpel-reqts] some ideas for our call today
Importance: High


I've been playing around with the categories from both Monica and the original
authors to explore possible discussion blocks for the meeting next week.
 I basically used venn diagrams rather than specific issue content,
level of interest shown in the TC to date, readiness for discussion, someone
to champion, etc. all of which may be more important in deciding what winds
up on the agenda.  So, given that it is a very rough view, I offer
it in case its helpful. 
I apologize for the lateness - I would have sent it sooner, but didn't
get do it till late last night and felt it needed a review during daylight
hours before sharing.   

Note, these aren't in any special order:  
    
1.  BTM:  53, 54, 55, 56, 57, 58, 59 (all opened by Choreology).

2. compensation related issues 
Error handling including compensation and permeability (authors categories):
1,3,10,25,27,30 
Compensation (MM categories):10, 20 25, 30 
State management - error handling (MM categories): 1, 3, 21, 27

3. Correlation and activity structures 
Correlation (MM) 8, 14, 26, 28, 29, 31, 32, 33, 37, 44, 49, 50
Activity structures (authors) 4, 14, 23, 28,44 (14 and 28 added since original
authors  input) 
Correlation (authors) 8, 33, 37 
 
4. Event Handlers (authors) 32, 36, 41, 43 (note 32 in MM correlation group
also) 
  
5. Correlation and async. operations (MM): 16, 17, 22, 23, 36  

6. State management activites (MM) 2, 4, 5, 6, 43 -  no comparable
group from authors (note issue 4 is also in authors activity structures)

7. Modelling/formalism (MM) :  18, 35, 38, 42 - no comparable group
from authors 

8.  
WSDL (authors): 7, 12, 15, 47; 
XML (authors): 11, 13, 29, 39, 46 
Technology (MM): 7, 12, 13, 15, 19, 40, 46, 47, 48
Assignment (MM): 11, 39, 51   
  
Regards, Diane
IBM  Dynamic e-business Technologies
drj@us.ibm.com
(919)254-7221 or 8-444-7221, Mobile: 919-624-5123