*MACE-Dir Conference Call*
May 14, 2001

*Attendees*

Bob Morgan (acting chair) - Washington
Tom Barton - Memphis
Steven Carmody - Brown
Tom Fowles - Penn State
Tom Dopirak - CMU
Todd Piket - Michigan Tech
Nate Klingenstein (scribe) - Internet2

*Discussion*

The call had difficulty meeting its agenda due to a lack of participants.
Keith Hazelton, the group's chair, was in Turkey and unable to dial in at
the last moment, leaving Bob Morgan to step in as chair.  The first three
agenda items had no representatives on the call, so the group addressed the
action items, then progressed directly to the fourth.  In light of the
action items not being posted far in advance of the call, the first
suggestion to arise from their review was to post action items within 48
hours, perhaps before the minutes themselves were written.

The only action item that could be reviewed was the third action item
regarding writeups of schools' experiences with groups.  It was unclear to
some as to precisely what should be detailed in these writeups. [AI] Tom
Barton volunteered to write a message detailing the information that would
be useful to have.  This would primarily discuss the various schema schools
have designed for their purposes, mostly based on a relationship between what
kind of object is represented by the groups and the applications relying on
them.  At Brown, for example, information in groups is used in both
communication and authorization applications.  Bob Morgan, formerly
of Stanford, was asked about the progress of that institution's
unconventional approach to groups.  Their system is based upon registries,
which each contain data for one principal, such as an application,
organization or a person.  To manage these registries, group-style lists
are arising in current implementations of this system.  More information on
this system is available at
http://www.stanford.edu/group/itss-ccs/project/registry/.  

Discussion next focused on the fourth agenda item, on groups, 
attributes, and roles. Bob's understanding of the current 
proceedings on groups, attributes, and
roles is that the group had been assessing Todd's writeup on his method
for groups and there was some enthusiasm.  Review of this document would be
helped by analysis of the aforementioned writeups.  Additionally, a
document written by Michael Gettes of Georgetown was forwarded to the list
immediately prior to the call.  Bob's initial understanding was that this
was a proposal for a specific means of linking directory entries among
multiple directories referring to the same object, known as stitched
directories.  There are some current implementations where information is
assembled from queries to assorted databases to supply the necessary data
to relying applications.  In one such makeshift design at Washington, no
administrative boundaries are crossed.  Permissions could prove difficult
in multiple lookups across organizations or institutions.

The necessity for such an application arises in two complimentary
situations, as detailed by Bob.  A large central directory may not want to
contain some attributes it would regard as peripheral, preferring instead
to refer the lookup to a local directory.  Michael's Directory of
Directories (DoD) is one example of such an application.  From the other
end, a small directory might contain the vast majority of a person's
information, but an application may want to call an attribute such as
central email address from a major database.

An obvious application that could benefit from a system such as stitched
directories is the Grid.  It was noted on the call that MACE-Dir tends to
raise this issue more than the Grid people themselves, which may be
telling. [AI]  People will read through Michael's stitched directory
document thinking of the problem to be solved and how well this proposal
addresses it.

*Action Items*

[AI]  Tom Barton will write a message detailing the information that should
be included in writeups of schools' experiences with groups.

[AI]  People will read through Michael's stitched directory document
thinking of the problem to be solved and how well this proposal addresses
it.