*MACE conference call, August 28, 2000*

*Attendees*

Bob Morgan (chair)
Steve Carmody
Keith Hazelton
Eric Norman
Ken Klingenstein
Renee Frost
Paul Hill
Michael Gettes
Neal McBurnett
Mark Poepping
Ben Chinowsky (scribe)

*Discussion*

The meeting opened with a review and critique of the nuts and bolts of MACE operations. There was general agreement that the current biweekly schedule for MACE conference calls should be retained; it was also agreed that the group would make an effort to keep the calls to a length of one hour, starting with this one. The Wiki Wiki Web -- a consensus-generating tool that works via shared editing of Web pages -- has been suggested for MACE's use; Bob thinks it would be worth a try, and asked for input. MACE expressed both interest in this technology and concern about some of its limitations, such as the danger of having your changes stomped on and many variants' near-total absence of security. Ben noted that the Internet2 QBone site is using Slash (the now open-sourced code behind Slashdot), and suggested that MACE consult Ben Teitelbaum for insights. Ken noted that there is broad interest in Internet2 in using Wiki-like tools, and that using Slash would have the significant advantage of "conservation of tools". [AI] Anyone who has pointers to good software of the Wiki/Slash ilk should forward them to MACE.

The results of voting for new MACE members were announced; David Wasley and Jim Jokl are the winners. [AI] Ken will send out a last call offering MACE members another day or two to provide input on new MACE membership. [AI] Assuming no objections, Bob will ask David and Jim to join MACE.

The discussion then moved on to project status updates. Ken introduced his proposal for the next stage of the directory-of-directories project, asking what review MACE should do before MACE-Dir takes up this project. There was general agreement that "buy-in is more important than review here", and that with Michael's continued commitment, Roland Hedburg's offer of participation, possible authorization work, and possible vendor interest, this effort is in good shape. Ken also noted that a recent external review of CISE recommends a major middleware initiative without mentioning Internet2 at all.

In the Shibboleth work, Bob noted that there had recently been a couple of useful phone discussions with IBM; they are looking at Web single sign-on products and working to produce a recommendation on which of them would be best for the project to standardize on. Ken has set up another phone call for Monday, September 11 for IBM to report back; this call is open to the MACE list. The Shibboleth group still aims to have something to report by the Fall Internet2 Member Meeting. Ken noted that there is interest in Internet2 in combining the file-discovery process of Napster with authentication, for use by K-12 schools. Ken asked MACE if they thought Shibboleth could support this. No consensus was reached on this issue; it was suggested that the directory of directories project might be a better home for such work.

EduPerson v0.9 has been released and is now under review by several groups. Ken asked what happens when the NIHPerson attribute definitions collide with eduPerson's; Keith informed the group that there are plans to meet with the NIH people, hammer out the differences, and get them into an RFC. NIHPerson appears to be "more cooked" than eduPerson, so it will likely be eduPerson that gives way. It was agreed that as lots of *Persons (e.g. gridPerson) are coming soon, it is important to develop a process for resolving such conflicts. [AI] Keith will compare NIHPerson and eduPerson and summarize for MACE-Dir; this is second in priority only to fixing the EPPN language. The institutional focus of NIHPerson contrasts with the inter-institutional focus of eduPerson; this led to disagreement about where (LDAP Recipe or elsewhere) to put the discussion of NIHPerson. [AI] Keith and Michael will discuss where to document issues around NIHPerson and other institutional *Persons, with a view to informing MACE-Dir's decision on this question. It was also agreed not to roll Ken's OIDs one-pager into the LDAP Recipe, but instead to post it separately and have the Recipe refer to it. Ken has been told that many of those present at Snowmass are impatient with I2-MI's PKI work and think that higher education should just outsource and use a corporate CP. While Ken is skeptical that this is really the case, he is concerned, and asked MACE for their impressions. There was general agreement that while people are impatient to see progress, there is in fact no widespread "just sign here" attitude among the Snowmass attendees or the higher education community more generally. It was observed that no one other than HEPKI and the Feds is focusing on inter-institutional PKI. MACE also noted an emerging consensus that liability is the central challenge of PKI; at the same time, vendors are generally very reluctant to accept liability, raising the question of what exactly those who choose to outsource will be getting for their money.

Ken noted that the Internet2 Fall Member Meeting would include medical-middleware sessions, and that both main-campus and medical-center IT architects would be attending. [AI] Ken will inform MACE of the specifics of a dinner planned for discussion of the medical-middleware work. [AI] Keith will discuss with Ken a medical-middleware-related grant opportunity at the University of Wisconsin.

The final topic discussed was the open-sourcing of AFS (see http://www.transarc.com/News/press/opensource.html). Should Internet2 be involved? There was general agreement that this is a promising development -- AFS is superior to NFS, and its limited commercial success has more to do with poor marketing and poor integration of suggested improvements than with any fundamental technical problems. Ken noted that he gets a lot of questions about whether Internet2 will be doing work with distributed file systems; it would be good if we could recommend something in this area. [AI] Mark will take charge of exploring the possibility of Internet2 involvement with AFS.

*Action Items*

[AI] Anyone who has pointers to good software of the Wiki/Slash ilk should forward them to MACE.
[AI] Ken will send out a last call offering MACE members another day or two to provide input on new MACE membership.
[AI] Assuming no objections, Bob will ask David and Jim to join MACE.
[AI] Keith will compare NIHPerson and eduPerson and summarize for MACE-Dir; this is second in priority only to fixing the EPPN language.
[AI] Keith and Michael will discuss where to document issues around NIHPerson and other institutional *Persons, with a view to informing MACE-Dir's decision on this question.
[AI] Ken will inform MACE of the specifics of a dinner planned for discussion of the medical-middleware work.
[AI] Keith will discuss with Ken a medical-middleware-related grant opportunity at the University of Wisconsin.
[AI] Mark will take charge of exploring the possibility of Internet2 involvement with AFS.