OMG Internet SIG - Minutes of Meeting #26 (original) (raw)

Minutes of Meeting #26

Mesa AZ

January 10-11, 2000

internet/00-01-01

co-chairs: Craig Thompsonand Shel Sutton

Attendance

Bill Cox - Novell - bill@novell.com
Edward Feustel - Institute for Defense Analyses - efeustel@ida.org
Jeffrey Kurtz - MITRE/Open Systems Center - jkurtz@mitre.org
Eric Newcomer - IONA - eric.newcomer@iona.com
Henry Rothkopf - MITRE/Open Systems Center - henryr@mitre.org
Sebastian Staamann - XTRADYNE Technologies AG - sebastian.staamann@xtradyne.de
Shel Sutton - MITRE/Open Systems Center - shel@mitre.org
Craig Thompson - Object Services and Consulting - thompson@objs.com
Robert Whitney - The Bulldog Group - rwhitney@bulldog.com

Minutes

� by Craig Thompson

Internet SIG Directions, Craig Thompson, ISIG co-chair

We had a brief discussion on the need for new directions for Internet SIG and decided to hold a brainstorming session at the Denver meeting and to include a review of the results of the Internet Services RFI as a started set of possible new actions. Many of thoserecommendationshave since been adopted but many have not so let's triage these and see what else is still needed.

Computer Supported Cooperative Work (CSCW) Working Group, Henry Rothkopf, CSCW Chair

A version 2 of the CSCW green paper (internet/00-01-02) is now available. It drops away some DoD references and uses an OMG example. Jeff Kurtz (MITRE) led a review of this green paper.

Critical things to do by next meeting

determine relationship to Task and Session, Workflow, Negotiation, ECTF Collaboration, and maybe Knowledge Management especially. Enterprise Application Integration SIG.
determine if and when we will do a CSCW RFP - complete a draft by Denver focusing on section 6 using new standard rfp template Other important things to do
explicitly separate collaboration meaning from UML collaboration diagrams, EC collaboration, and CSCW collaboration
consider interoperability among collaborative workspaces but maybe explain why not to consider it in an RFP
should CSCW stay under Internet SIG or move nearer to workflow and task and session
areas like role and policy are under specified

Discussion on Issues

Policy discussion notes. There are many broad interests in OMG onpolicy. One short description - policy is the behavior of the system as supported by the system developer and controlled by the owner. Policies for allocation and shared resources are at server and behavior of how to display to user is at client. (But not always).

Related work on policies: Morris Slowman in England has been working on policy languages. Are policies consistent. Security policy languages like Detail++. Work by Greg S, Doug Maughm. Policy Languages as in Polar Sudan's. Telco is working on network management. Paul Kaitv from Rogue Wave.

OMG needs a general Policy Management Green Paper. How broad a framework do you want? Enterprise uses. Perceived issues of control and management. Policy is a statement of business rules. It provides roles and responsibility, rules, enforcement, authority to do X on data Y. Define the set of roles and their duties and what sorts of framework variables can be controlled (by mechanisms that support the policy choices). What are the administrative tasks. Use cases of administration of these activities. Must be dynamic because people's roles change on an immediate basis. One thesis is to allocate tasks by role and when you start doing a task then you get a new role. Dynamic assignment and reassignment.

Composing in other Services. How to add the CSCW framework into workflow, security, e-commerce, and task and session. How does QoS fit?

workflow and task and session are (perhaps) super and sub sets of CSCW and we want some interconsistency
security can be interlaced in several ways, via security unaware or where each object is supported. What kind and when do you need non-repudiation. Possibly many ways to do this.
not sure how EC negotiation fits except that collaboration participants might want to negotiate among themselves or possibly with outsiders. And the EC may have frameworks for people and sessions that are relevant. Observation: OMG does not have a good language and set of interfaces for fully describing composition in terms of being able to introspect on interfaces and implementations to see how a thing is composed and how to replace modules or insert new behaviors. This is a big hole.