Meeting Notes: Review of Repository Information Model v0.1

[Farrukh Najmi <Farrukh.Najmi@east.sun.com>]

Below are the meeting notes from the review. I felt encouraged that it
was an excellent and very productive meeting. Go team!

Lets do it again tomorrow at 1pm EDT for the RS spec.

Attendees
------------

Scott Nieman, Yuta, Scott Hinkleman, Lisa, Joel, Len, Farrukh

Issues Identified
-------------------
This is a best effort. If I missed one please post correction. I will
either address them, stick to my position or include them in issues
section of v0.2.

Several good issues from Len's email on list not included here.

Terminology Mapping Table
---------------------------------
Need forward ref so people know it is there

Managed Object Confusion
--------------------------------
There was a general confusion about the use of the term ManagedObject to
refer to meta data about the actual  managed object (e.g. DTD) vs. the
actual managed object itself. Many suggestions were made. I will post
some suggestions and see what you like.

Yuta's Alignement comments
-----------------------------------
Need to address terminology alignment comments that Yuta had
painstakingly identified in v0.4. Note this was done about 50%. Yuta
your comments where on the RS spec not RIM spec. I will try and
extrapolate.

Discussion on Object Class Hieracrchy
----------------------------------------------
-Object is too generic (SH)
-Name choice for ExternalObject (Len?)
-Associations questioned between ManagedObject and Object. I believe
this issue was resolved to leave it the way it is/


globalName managed object attribute
-------------------------------------------
Farrukh raised the issue that globalName managed object attribute (from
OASIS) is not a good idea. Len explained the intent was to uniquely
identify an object in any registry with a user friendly name. Farrukh
responded that uniqueness is the role of a GUID while userfriendly name
need not be unique. Farrukh mentioned that this is a fairly well
established distinction and that ebXML arch team wants to
instititionalize GUIDs in ebXML. Farrukh mentioned that the id attribute
in ManagedObject was being changed to guid to make it obvious that it
was a GUID based on arch teams comment. I dont think teh issue was
resolved but I feel pretty strongly on this point.

Figure 2
----------
-Missing Association class box on association between ManagedObject and
ExternalObject (LC)
-Need a separate Association class for association between managed
object and managed objects (LG). I feel strongly that I have not   heard
a good reason articulated (no change planned for now)
-Support MixIns in model (SH). farrukh professed ignorance. Will talk to
SH offline to get educated.
-Classification should not be a ManagedObject (LG). I have a strong
opinion that anything that is submitted is a ManagedObject (no change
planned for now)

Object Life Cycle Issues
------------------------------
-Lisa mentioned that we need to support the distinction between an
object that has been submitted but not yet managed. Defer for Tokyo.
-Len mentionedthat a submission could be rejected. Farrukh said the
entire ObjectManager workflow is inadequate for the long term and what
we have is stream lined for Tokyo.

Organization: Sub-class Vs. Roles
----------------------------------------
Len mentioned that SubmittingOrganization etc were roles of an
organization and not sub-classes. Farrukh agreed whole heartedly since
the same org could be a SubmittingOrganization and a
ResponsibleOrganization in same or different contexts. Will need to make
sure that BB/CC align with our thinking. farrukh to take it up with
them.

Object Versioning
---------------------
Lisa: may need to support Submitting Org defined versions in addition to
the meta data support for version that are assigned by RS. Defer for
Tokyo.

Object Association
----------------------
Len: Objected to association between ManagedObject and Object. Farrukh
feel sthis is OK
Len: Need to support constraint rules (targetConstrainst) attribute in
Associations (e.g. OCL). Include in isses for post Tokyo
Scott N: Need better semantically equivalent example than Fig 5. Farrukh
asked team to provide some examples.

Package Support
--------------------
Lisa: need it. Farrukh agree. Post Tokyo

Figure 7: Classification stuff
---------------------------------
Len: Missing a link from ClassificationScheme to ClassificationItem.
Farrukh felt that while the OASIS model was good
it could be improved further. Scott and Farrukh to work offline to
improve.

Registry-registry object traversal
--------------------------------------
Len: farrukh will include in issues for post Tokyo

Required column on attribute table
-----------------------------------------

-Need to clarify who sets the required value client or RS. Farrukh
agreed










--
Regards,
Farrukh

begin:vcard 
n:Najmi;Farrukh 
tel;fax:781-442-1610
tel;work:781-442-0703
x-mozilla-html:TRUE
url:www.sun.com
org:Sun Microsystems;Java Software
adr:;;1 Network Drive, MS BUR02-302;Burlington;MA;01803-0902;USA
version:2.1
email;internet:najmi@east.sun.com
fn:Farrukh S. Najmi
end:vcard