*MACE Conference Call*
October 21, 2002

*Attendees*

Bob Morgan (chair) - Washington
Renee Frost - Michigan/Internet2
Paul Hill - MIT
Von Welch - Chicago/Argonne/NCSA
Diego Lopez - RedIRIS
Michael Gettes - Georgetown
Steven Carmody - Brown
Jim Jokl - Virginia
Scott Cantor - OSU
Ken Klingenstein - Colorado/Internet2
Steve Olshansky - Internet2
Keith Hazelton - Wisconsin
Neal McBurnett - Internet2
Ben Chinowsky (scribe) - Internet2

*Discussion*

The meeting opened with a discussion of the increasing interest information architects from the private sector (Microsoft and Sun in particular) are displaying in the work of the Internet2 Middleware Initiative. Some of these individuals will be at the Internet2 Member Meeting next week; fruitful discussion is anticipated. Ken also asked the group to give some thought to how best to "take advantage of the commercial mindshare we seem to be gaining" in the months following the meeting.

Keith noted that, based on comments from NMI testbed sites, MACE-Dir considers the revised versions of the NMI-R1 metadirectories-practices and groups-practices documents to be ready for release. Keith asked for MACE's go-ahead, which was duly given.

The group discussed the recent Global Grid Forum meeting. In the Grid community as elsewhere, authorization issues are increasingly the focus of attention. Von noted disagreement between those who want to wait to see what directions IBM and Microsoft take, and those who want the Grid community to go forward now. [AI] Von will send MACE some comments on what code will be needed for authZ efforts in GGF. Bob observed that there are a great variety of independent plans in this area; it seems as though each vendor that doesn't have its name on the WSS roadmap has a roadmap of its own to promote.

An Internet2-wide consensus is beginning to emerge on the document-standards document; Ken's sense is that there is unlikely to be much dissent from MACE's final decisions on this. [AI] All will send MACE their comments on Nate's latest version of the document-standards document (see attachment to Ben's email of Friday, October 18.)

Diego noted that there's a whole new crop of middleware projects getting underway in various places, and Ken noted that NMI is opening up to include a wider range of types of contribution from a wider variety of sources, especially international sources. Together these facts mean that deciding what to include in NMI-R3 is much more likely to involve selecting from a surplus of candidate components, than coping with a scarcity of such components. Ken suggested that discussion of this decision-making process be organized on two levels: a more general level, addressing questions like "What contributions should NMI accept?" and "Who should pay for integration?"; and a more detailed level, addressing questions like "Should all code releases be required to use Bugzilla?" Ken observed that Grid software tends to be fairly tightly coupled; this makes it easier for the Grid community to decide what to include, but harder for them to accept things not specifically developed for the Grid. Keith noted that while political considerations make people want to include things, "ad hoc-ness is the ultimate enemy"; clear criteria for inclusion -- e.g., "Is this a component that fosters integration?" -- need to be developed in order to avoid ad hoc-ness. With NMI-R3 planned for next April, there was general agreement that getting this discussion started is fairly urgent. [AI] In order to start discussion of what the criteria should be for inclusion of items in NMI-R3, Ken will send MACE a list of items that could possibly be included. Diego noted that the Stockholm TF-AACE meeting will include representatives of most of the European middleware community, making it a good place to raise these issues.

Finally, Bob noted that the DRM working group now has a draft charter under review, and Ken noted that this group will be a CNI/Internet2 joint effort, on the model of the VidMid ViDe/Internet2 joint effort.

*Action Items*

[AI] Von will send MACE some comments on what code will be needed for authZ efforts in GGF.
[AI] All will send MACE their comments on Nate's latest version of the document-standards document (see attachment to Ben's email of Friday, October 18.)
[AI] In order to start discussion of what the criteria should be for inclusion of items in NMI-R3, Ken will send MACE a list of items that could possibly be included.