[16] in Kakapo Windows Team
Minutes of the June 19, 2003 Kakapo Meeting
daemon@ATHENA.MIT.EDU (Kerem B Limon)
Mon Jun 30 20:09:49 2003
Message-Id: <5.2.0.9.2.20030630183252.07b366e8@hesiod>
Date: Mon, 30 Jun 2003 20:09:31 -0400
To: Kakapos at MIT <kakapo@mit.edu>
From: Kerem B Limon <kerem.limon@MIT.EDU>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Enclosed are my meeting minutes for the June 19, 2003 (first) Kakapo
meeting. Please send all corrections to me and I will issue an updated
version if necessary. My apologies for the delay in getting this out.
Regards,
Kerem
---Enclosure:---
Kakapo Meeting Minutes
Date & Time: Thu, June 19th, 2003 @ 1300 - 1430 (1:00 PM - 2:30 PM)
Location: W20-PDR4
Attending: Tom Coveney, Chad Dupuis, Jean Foster, Paul Hill, Jonathan
Hunt, Kerem Limon, Phil Long, Brian Murphy, Oliver Thomas, Tom Thornton,
Jonathon Weiss
Absent: None
1) Introduction (Phil) - item #1 on Brian's agenda
Phil mentioned that most in the room have heard by now about this group and
why we were meeting. He thought that probably most all in IS do not deny
the need to coordinate Windows efforts around the Campus. He mentioned that
one model suggested is the Athena Owls model. Phil reminded, though, that
we do not either want to impinge on or ride off the back off Owls too much,
hence the new group.
Phil explained that a kakapo is a chiefly nocturnal Southeast Asian/New
Zealander bird, which for all intents and purposes is an owl, but is
colored green. It's known for its mating call which is loud enough to be
heard over very long jungle distances. (For those not in the know, Phil
happens to be an ornithologist in disguise :))
Phil raised the question of membership. He felt that this group in the room
was a good start, but that we should be more inclusive. He suggested it
might be useful to have a core group and a rotating larger membership like
Owls does.
The Kakapo group is co-sponsored by Phil and Theresa Regan to reflect the
cross-organizational and academic/administrative nature of the Windows work
needed.
Phil sees this group as "coordinating the delivery of Windows services for
deployments across campus". He mentioned that the directors and Jim (Bruce)
has been meeting every 3 weeks or so recently to come up with a reasonable,
supportable Windows delivery strategy. So this group has/will have a high
visibility in that arena.
Tom (Thornton) asked the question of availability of any funds,
particularly for certain hardware purchases like owls is entitled to. Phil
explained that this may be the case to the extent that owls does have it,
but it is more likely to come in the form of making recommendations as to
where and how to make hardware expenditures.
Brian asked the question of how owls became as influential as it currently
is; was it through active membership? Jonathon (Weiss) answered that over
time the Owls team developed this kind of influence, some through
representative membership from influential groups in IS. Phil added that
the people around this table is a good representation of that concept as
far as Windows at MIT is concerned.
2) Charter & the Team (Brian, open discussion) - items #2, #3 on Brian's agenda
Brian asked for suggestions for names to approach for membership,
particularly non-IS folks.
Kerem suggested Mike Sherman from Facilities and Steve Dowdy from OSP,
noting their long-standing involvement with our past and current Windows
Server efforts. He cautioned however, to be clear about what the commitment
to Kakapo entails and what they would be expected to; otherwise they may
feel they are wasting their time on what may seem to them like yet another
IS group. Tom (Thornton) seconded Kerem's comments.
Tom (Thornton) went on to suggest some container admins, for instance,
Pat(rick) Paul from Bio/Micro.
Paul suggested also being careful about the time commitments we ask for,
especially in cases like Pat. He noted that with IS not having provided a
lot of support for research, we should be careful about the level at which
we engage them.
Paul added that there is also the container admins' monthly meeting which
can be used as a forum to reach out to them.
Brian mentioned Jed (lastname?) from Steve Lerman's area.
Phil suggested Tom Fitzgerald from Architecture. Chad countered saying that
he was not sure as to his potential time commitment, having observed them
(inc. DUSP) not even attending container admin meetings.
Tom (Thornton) asked Brian and Phil if they had thought of possible student
representation. He wondered how Owls deals with this, particularly the
difficulty with students being unable to make long term commitments given
their academic priorities. Oliver added that we want to be careful about
"including input from students" rather than "having a token student" and
disinteresting them repeatedly with mostly administrative agendas, should
that be the case from time to time. Jonathan (Hunt) emphasized, though,
that we should keep the students in mind in our meetings, especially one of
their issues are up for discussion. Oliver replied that he would be OK with
having a short list of students on the standing membership list, but would
rather invite focus groups representing those contingents that would be
affected by the decisions.
Jonathon (Weiss) commented, in line with student interaction, that it would
be useful to make the list be a public distribution list.
Jonathan (Hunt) asked about including someone from the Help Desk. Kerem
seconded his recommendation. Kerem also raised the issue of other
front-line teams beside the Help Desk, like BLT and Athena Consulting who
may potentially be impacted by Windows-related changes. He suggested Fred
Baars from the Computing Help Desk as a potential member.
Oliver went on to suggest that it may be a good idea to include the team
leaders of participants in the mailing list even if they send a proxy to
the actual meetings, just to avoid any misunderstandings as to commitments
being made on behalf of the participating groups and/or the decisions being
made getting communicated accurately to these groups.
3) Short-term Issues (roundtable) - item #4 on Brian's agenda
3a) Request for MATLAB GIS Toolbox installation and other software in
Building 37 Cluster (items #4A and #4B on Brian's agenda)
Oliver expressed concern that both of these are tasks that steal from
Chad's time. He also mentioned that we'd probably like to make the
installation consistent so that it may run in the Libraries GIS environment.
Paul suggested that perhaps running the applications off of AFS may be
possible. Chad commented that the license management mechanism these
products use, breaks that.
Phil commented that cost is not a major concern given the numbers, but he
wanted to see if the Libraries had a better environment for dealing with
this. Oliver Thomas will check with Daniel Sheehan and see if that's a
possibility. Otherwise we have a consensus to go ahead and make this happen.
3b) Keyserver services sought by Architecture/DUSP (item #4C on Brian's agenda)
Oliver explained that the new software requests in question are coming from
Jed. There is no new software, hence costs here, the licenses are in place,
but there is the overhead/cost of installing them on the machines.
Paul suggested, if necessary, bringing back the Ken Hilker, the MSI
consultant from Wise Solutions. Chad commented that he was not sure that
was necessary, and that he would think about it. Phil also observed that
this relates to our MSI development strategy. He noted, however, that
separately, dealing with keyservers, the question is whether we (as IS)
need to provide keyserver services Campus/domain-wide. He went on to say
there may be a need for a mini-Discovery type project or small such effort.
Paul commented that he would like to tie this into our MSI efforts to, for
example, accommodate proper keyserver configuration/selection implemented
through MSI transforms.
Oliver noted that the software Architecture/DUSP use can accommodate
multiple servers and refer to one another. He suggested that in the future,
we could have central and delegated key servers. Phil added that we could
possible leverage key serving over the same servers to other departments,
assuming the licenses are verified to have been acquired and properly
installed.
Jonathan (Hunt) added that they (the Software Release Team) was trying to
move their Windows installers over to MSI versions, especially in the newer
efforts. Atticus (Gifford, the installer writer for SwRT) has been working
with the win.mit.edu folks on this. Chad added that in most applications,
we may need two versions based off of some common root, as stand-alone,
customer/end-user-oriented versions of installers may not immediately work
being deployed off of Group Policy. Kerem seconded this, citing previous
conversations with Chad and Atticus on the same topic. Jonathan (Hunt)
noted that win.mit.edu has a much larger suite of software than those
distributed through Software Release.
Phil wrapped up by observing that overall we were all comfortable with the
keyserver discussion and the work in this area should go ahead as requested.
3d) MSI installers development coordination (item #4D on Brian's agenda)
Phil opened the topic by noting that a lot of separate efforts may not be
so efficient. Paul noted that while some vendors ship products already with
MSI installers, others don't and some departments than end up having to
write their own to meet the need. Paul added that there is obviously more
than Software Release could alone handle, so for the we may not able to
standardize across various efforts in the foreseeable future so easily.
Phil and Paul, however agreed that some guidelines would be good. Chad
commented that he has done some documentation as part of his work in
Academic Computing and this documentation should be coming up soon.
Paul raised the suggestion of keeping the list of installers available as
perhaps an XML document and than being able to transform this easily into a
subset that can be presented depending on the context (e.g. 'academic'
applications, 'business' applications, etc.). He further suggested CVS
(Concurrent Versioning System) for keeping track of the edits to that
document. Jonathan (Hunt) wondered how this would fit with the stand-alone
installer information, while Kerem explained Paul's suggestion in a bit
more detail. Paul added that style sheets can be created to extract the
needed information from the XML document. Jean commented that customers may
not necessarily care about where the document is or from what team it
comes. Kerem further explained that Paul's suggestion means maintaining
simply one (1) document and we can each do appropriate extractions of
subsets of information from that for different venues, team-oriented or
not. Jean then asked the question of who would maintain said document. Paul
suggested that once we have an agreed-upon schema, we would give access to
whomever is authorized to develop and release these installers. Jonathan
(Hunt)added that CVS would then accommodate concurrent editing. Paul
suggested it could be branched off of the win.mit.edu CVS tree. Kerem
suggested that Tom and he could discuss details since he is already
involved in doing similar work for documentation through his Windows
Servers Delivery Team, but that someone would need to draft the schema.
Paul commented that Joe (Calzaretta) could probably do this, but warned
that it may take a month given the number of things he is currently working on.
3d) Other issues and ongoing work (item #4E on Brian's agenda)
Kerem gave an overview of the ongoing Windows Servers Delivery work. He
explained that we are working to launch a volunteer consulting team called
the Design Assistance & Review Team (DART) composed of IS personnel
involved in Windows and knowledgeable Windows admins from the departments,
labs, and centers (DLCs). He explained that the role of this group will be
help assist DLCs migrating from Windows NT 4.0 to Windows 2000 in choosing
whether they should join win.mit.edu or deploy an independent Windows 2000
Active Directory (AD) domain. The process involves a per-DLC case-manager
assigned from within DART, who sets up a number of negotiations with the
DLC to gather the pertinent information, understand the needs, identify
matches with either option, and make recommendations. DART then hands over
the department gracefully to resources identified for actually carrying out
the migration work--these can be the DLC's own resources, DITR, or even
Microsoft Consulting Services (MCS) or Dell Professional Services (DPS).
DART includes people from DITR and we have worked with MCS and DPS during
the Delivery Project phase for some early adopters.
Kerem then gave a list of the departments currently engaged in migrations
or have completed them as follows, along with primary contacts/case managers:
o Architecture - in migration to win.mit.edu (Chad)
o Audit & Budget - potential win.mit.edu customers, with laptop-related
issues awaiting clarification (Tom (Coveney), Kyle Pope)
o Controller's Accounting Office (CAO) - test-ran the first phase of DART,
interested in win.mit.edu, report just out through Mike Sherman (Chris
Meehan, Mike Sherman)
o (existing) DITR customers - in planning (Tom (Coveney), Kyle Pope))
o Facilities - planning migration to independent domain due to strict
business needs, including but not limited to Exchange Server (Mike Sherman
and DPS)
o Information Systems - internal migration of groups using Windows NT 4.0
domains ramping up (Tom (Coveney), Jonathan (Hunt), Kerem, Barry Stoelzel)
o Libraries - currently testing win.mit.edu, inclined to proceed joining
(Kerem, Wael Hishmeh)
o Nuclear Reactor Lab (NRL) - having tested win.mit.edu, moving towards and
independent AD domain (Ed Block)
o OpenCourseWare (OCW) - implemented an independent Windows 2000 AD domain
due to compatibility needs with outside contractor, Sapient (Kerem)
o Professional Learning Center - migrated administrative desktops and
server to join win.mit.edu, migrated clasrrom environment desktops and
servers to independent AD domain to meet teaching needs (Steve Goldman, Tom
(Coveney), Wael Hishmeh, Kyle Pope for win.mit.edu migration, Steve Goldman
and DPS for independent domain)
o Sponsored Programs (OSP) - migrated desktops and servers to win.mit.edu
(Steve Dowdy)
o Treasurer - deployed independent AD domain to meet business needs,
particularly Exchange Server in support of Blackberry RIMs (a number of IS
folks, inc. Theresa, Carol Wood, Wael Hishmeh, Paul Hill, etc. at various
stages, and MCS)
o Urban Studies & Planning (DUSP) - in migration to win.mit.edu,
collaborating with Architecture (Chad)
Kerem raised the question of what level of nesting of containers is
allowed/recommended in win.mit.edu, now that this is a new feature
available. He was specifically concerned about the support
statement/implications around this recommendation, which he has casually
been told to be at most 2 levels deep. Paul commented that the support
status of this offering is the same "as much as anything is supported in
win.mit.edu". Kerem asked to define a clearer description in a future meeting.
Paul commented that Wael Hishmeh will be transitioning to DITR at the end
of the fiscal year. Kerem mentioned that he had had conversations with Wael
about how to keep him engaged in the migrations during his transition and
after, and that he'd be following up with Kyle Pope, Tom (Coveney) and Tom
(Thornton).
Phil mentioned that for starters, DITR would perhaps offer some gratis
services to get customers going and/or to entice them to join win.mit.edu.
This reminded Tom (Thornton) that there is ongoing work for cost-modeling
services in the Windows arena.
4) Next Steps/Meeting Schedule (item #6 in Brian's agenda)
Kerem reminded that we will need a mailing list for this group and
preferably a Discuss archive. Brian agreed to create the list and Jonathon
(Weiss) and Oliver offered help with the Discuss archive through Accounts.
Brian will also look into this.
The time slot seemed reasonable to most people, and we decided to keep the
1300 - 1430 (1:00 PM - 2:30 PM) slot, with the intent to keep the
discussion within an hour when possible. We agreed to try meeting weekly
for a couple of weeks to come until some momentum is gained. There was
general agreement that this location (W20) was acceptable, but that
availability was not confirmed. Brian agreed to look into this.
We adjourned at approximately 1445 (2:45 PM).
Kerem B. Limon
Project Team Leader
Delivery Process, MIT Information Systems
kerem.limon@mit.edu /e-mail
+1-617-253-1150 /voice
+1-617-258-8736 /fax
