ebXML Registry and Repository teleconference minutes Thursday September 13, 2000

["Nieman, Scott" <Scott.Nieman@NorstanConsulting.com>]

The ebXML Registry and Repository project team held a teleconference
Thursday September 13, 2000.  
It lasted from 9:00a.m. to 10:40 a.m CDT.

Attendees:
Frank Olken, Lawrence Berkeley National Laboratory
Scott Nieman, Norstan Consulting
Farrukh Najmi, Sun
Lisa Carnahan, NIST
Joel Munter, Intel
David Webber, XMLGlobal
Bruce Peat, eProcess Solutions


=================================

Minutes:

Scott emailed Farruhk the mdl file he had of X3.285 for his review.  Scott
strongly felt that the classification meta model section deserved
significant review.  Frank Olken joined the conversation and provided to the
listserv the URL for the most current doc and pdf versions of X3.285.  Frank
stated that he was busy drafting the US position of X3.285 for the addition
to ISO 11179.  It was agreed that X3.285 would be reviewed; significant
areas for implementation were that of data element concepts and
classification.  Scott stated that he believed that the model could be
abstracted to be any type of submitted object rather than specifically data
elements, and recommended that viewpoint for our analysis and design.
Farrahk plans to submit to the listserv information likely in UML form
discussing the association of concepts to classifications.  Frank pointed
out the data_element_concept class in X3.285; from this it appears that
concepts and classifications are related by containment in that model.
(Scott encourages further discussion on the listserv; Frank, could you
elaborate on this?).  

Bottomline:  ebXML rep-rep will attempt to align as much as possible to
X3.285 for its information metamodel.

Since v0.5 was not available for review (Farruhk did not post since he was
waiting for Scott's preview) the team decided to review Farrukh's
requirements statements previously posted to the listserv.  David Webber
joined in at a time when TRP was being discussed.  The conversation quickly
became very heated and out of control.  Scott stated that TRP was a minimum
ebXML requirement (clarification post discussion:  this is primarily to show
the overall functionality of the ebXML architecture). David Webber felt it
implied asynchronous communication and Farrukh explained that TRP also
provides synchronous communication.


When the conversation moved to classification scheme, the conversation
totally degraded, especially when Farrukh suggested that classifications may
have an inheritance relationship to another.   David shifted the
conversation to his classification and interfacing document, and Farrukh
stated that the document was completed off the mark.  This got into a
discussion regarding OASIS specifications, including the notion of primary
and secondary classifications, as well as more discussion on X3.285.  Lisa
felt that David's document was dealing with some aspects that we would
eventually have to consider.  Scott stated that may be challenging to "meet
in the middle on this one".

We continued into the query sections.  Here, David felt Farrukh's
requirements were implementation specific.  Scott felt the requirements were
NOT implementation specific, and were common query concepts.  Scott also
felt an overall requirements-based approach followed that of a methodology.

Farruhk requested to Scott that he have the chance to be able to develop the
specification for managing classification schemes.  Scott stated that he
would make a decision by the end of the day.

********************************************

Scott's decision:

I was very disturbed by the meeting yesterday.  It was very unprofessional,
and I will take some of the blame for not shutting the meeting down earlier,
or asserting more control of the conversation.  It is not correct for
individuals to trample others statements.  

I am granting Farrukh his request to lead the development of the
classification scheme management aspects of the registry service
specification and ensure via his existing contact with the other project
teams that the classification requirements are defined and understood.   I
strongly believe that classification management needs to be specified in
UML, with serious consideration of X3.285.  Farrukh is very capable of
delivering this specification in UML and we have a commitment from Sun of
his focused involvement to this deliverable and the Tokyo POC.

The team will now only have one vice-lead, Yutaka Yoshida of Sun.

The original request to David Webber was that he manage amd coordinate the
classification scheme deliverables from the other project teams specifically
BP, CC and TP.  Unfortunately, David spent too much time developing a
document that included interfacing requirements which is the responsibility
of the interface team (Yuta, Farrukh et al).  The original task was to
assess the classifications from each project team and to ensure that the
query service Farrukh specified in UML was capable of handling these schemes
(no extraneous requirements).  It is very unfortunate that this document was
produced in this form, however, I am recommending that we ensure at minimum
sections 2.2 through 2.5 regarding classification attributes be considered.
(For that matter, we previously produced a document in Brussels regarding
classification attributes that also needs to be considered - see the reg-rep
private URL).  Other aspects of David's document may be considered later.  

I believe that we must continue the pace of the reg-rep effort that Farrukh
has contributed since his involvement.  We must continue down the UML path
and also begin the specification of a UML based registry information model.

I apologize to those who may be taken a back by this email.  If you were on
this teleconference, you would have been more taken a back, guaranteed.

*********************************************************


Respectfully submitted,  

Scott Nieman
ebXML Registry and Repository Project Lead