Policy Committee Meeting Summary: June 28-29, 1994
Boulder, Colorado
Participants
| Members | Representatives | UPC Staff |
| Robert Fox (Chair) | William Bonner (UCAR) | Sally Bates |
| Otis Brown | Bill Buzbee (NCAR/SCD) | Ben Domenico |
| Russell DeSouza | David Fulker (UPC) | Linda Henderson |
| Steven Mullen | Clifford Jacobs (NSF) | Linda Miller |
| Perry Samson | Mohan Ramamurthy (Users Committee) | Sandra Nilsson |
| Carlyle Wash | John Snow (UCAR Board of Trustees) |
| | Tim Spangler (COMET) |
Administrative Matters
- The summary of the March 1994 meeting was accepted as written.
Action 1:
Carry action items 1, 2, 5, and 8 from the March 1994 meeting over to the
October 1994 meeting.
Action 2:
Fox will issue a letter of invitation to Joe Friday and Dick Greenfield to
meet with the committee during the October PolCom meeting in Washington,
D.C. to discuss the status of international data commercialization and
exchange.
- The next meetings of the Policy Committee will be
6-7 October , Thursday and Friday, in Washington, D.C.
2-3 February 1995, Thursday and Friday, in Boulder, Colorado
- Prior to Policy Committee meetings, the committee chairman may, as his
needs dictate, hold an informal pre-meeting to facilitate the upcoming
meeting. These pre-meetings (zero, one, or several) are usually held the
evening prior to the Policy Committee meeting, with attendees as sought out
by the chairman. The pre-meetings are used by the chairman to better
understand the priority of issues and to be able to facilitate the meeting
itself in an efficient and expeditious manner. These pre-meetings are
mentioned here for clarity and openess; no policy decisions can be taken
during these pre-meetings because all policy decisions require action by the
full Policy Committee.
Action 3:
The Unidata charter may need to be stated in an independent document. The
statements from Chapter 5 of the Unidata proposal of May, 1992 to NSF were
distributed and reviewed. Fox and Bonner are to consider if and how to
clarify the Policy Committee charter and terms of reference, and will report
back at the next meeting.
Resolution 1:
The Policy Committee thanks John Nielsen-Gammon for his many contributions
as a member of the committee and looks forward to his continued involvement
in the Unidata Program.
Status Reports
Director's Report
Fulker's report is in the notebook; copies of his transparencies were
distributed at the meeting.
Discussion
- Fulker asked for Policy Committee feedback on the draft policy statement
on IDD principles. The topic was discussed again later in the meeting.
- Alden is willing to provide DIFAX to the community via the Internet by
December 1994. This agreement needs to be negotiated on behalf of the
community and the terms need to be defined. This issue was discussed in
detail later in the meeting.
- There was considerable discussion about obtaining COMET case studies and
images for the university community, as a result of the resolutions passed
by the Users workshop earlier in the month. Points raised during the
discussion included:
- formal letters to the head of NWS and NOAA should be reserved
for critical issues;
- UOP would be willing to look into the COMET case study question;
- the Users committee might need to form a working group to look
into what's really wanted;
- providing case studies is not a UPC function; it would make more
sense as a UCAR activity;
- workshop participants want more than simply case studies; what
they really want are materials that help them teach, which includes other
types of curriculum materials as well.
- Discussions of Alden's switch to Ku band were deferred to later in the
meeting.
- In response to questions about the delayed integration of netCDF into
GEMPAK and McIDAS, Fulker said that demands and opportunities for new
features in these applications, underestimation of the difficulty of
integration, and the redirection of effort toward IDD all contributed to the
delay.
Action 4:
The topic of Ynot will appear on the agenda for the October meeting.
Budget Report
Copies of Nilsson's transparencies are in the notebook.
Discussion
- The actual expenditures include money spent on the Users workshop;
Nilsson expects full reimbursement for these expenses from NSF.
- The transparencies do not show the impact from the new UOP overhead
rate, which has moved from 24.1 to 26.4%.
NOAA Report
Fox noted that he had participated in the AWIPS review panel and had signed
a nondisclosure agreement to do so. His NOAA report was therefore drawn
from materials provided by Doug Sargeant:
- The international commercialization of data is still a concern.
- Concerning AWIPS policies
- NWS access to NOAAPORT: anyone can have access as long as they
install the appropriate equipment.
- Changes to FOS: FOS will continue as long as there are
subscribers; there are no anticipated changes to FOS but FOS data will be
"neater" and easier to use than NOAAPORT data streams.
- NWS support for NOAAPORT users: NWS will not support NOAAPORT
end users; all users must purchase their own hardware and develop software
to assimilate the five data channels.
- Justification for end users expending money on increasing speeds
and volume of FOS if eclipsed by NOAAPORT: FOS may not be eclipsed and will
always be more "user friendly."
- NIDS: NWS does not anticipate an impact of NOAAPORT on the
current NIDS agreement. Although some NIDS data may move across NOAAPORT,
it will be in back-up mode or during extreme weather events.
- Schedule for AWIPS: No changes from the AWIPS published
schedules were acknowledged.
Discussion
- One member reported that Bob Landis of NWS had indicated that FOS would
be the only "certified" (i.e., quality-controlled) data stream.
NSF Report
Copies of Jacobs' transparencies were distributed at the meeting. Items of
note: in the future, funding for programs will be evaluated in terms of how
the programs fit within the strategic initiatives such as HPCC and IITA; the
Branscomb report has not been accepted officially.
Discussion
- The program to supply seismic data to K-12 schools predates the GLOBE
program and may be subsumed by it.
Users Committee Report
Copies of Ramamurthy's transparencies are in the notebook.
Discussion
- The smaller attendance of this summer's workshop (relative to the last
such event) reflects the more specialized nature of this workshop as well as
a substantial number of competing activities.
- Use of COMET case studies will require the UPC to alter its software
(i.e., it will depend on netCDF integration).
- The resolutions passed by the workshop participants involve COMET and
the UOP more than Unidata. COMET's mission is to focus on the needs of the
NWS; the UOP may need to help make COMET materials useful to the
universities.
- If the community wants action on case studies, it first should convince
the UCAR Member Representatives; if the MRs support the initiative, they can
ask UCAR to take it on. Passing a resolution in a vacuum, without
identifying a responsible institution or funding source is unlikely to be
effective.
- The focus of Unidata activities on the mesoscale was questioned. It was
argued that the Unidata community needed to be broadened to include the full
range of geosciences. Incorporating data streams such as EOS, UVB, drifting
buoy, and seismic data was mentioned. As IDD progresses, Unidata will find
the definition of high-interest data may broaden. However,
- Appointments to the Users Committee to date have resulted from
suggestions made by the committee and the community in general, with the
selection made by Fox in consultation with Fulker, Bonner, and the User
Committee chairman. In the future, this procedure will move to more open
discussion of candidates within the Policy Committee as a whole, with
selection the result of committee and UOP actions.
Action 5:
The UPC will encourage the subcommittee preparing the workshop report to
redraft the resolutions so that the efforts involve the UOP/COMET and the
universities.
Resolution 2:
Given the report on Unidata/COMET workshop entitled Teaching Mesoscale
Meteorology in the Age of the Modernized Weather Service, the Policy
Committee recommends that the UPC review the implementation schedule of
netCDF compatibility and report on updated schedules at the February 1995
meeting.
Resolution 3:
The Policy Committee commends the Unidata Users Committee and all the
sponsors for the creation of a very successful workshop.
Resolution 4:
The Policy Committee will seek to broaden the Users Committee to encompass
representatives from the Geosciences.
Internet Data Distribution
A copy of Domenico's presentation is in the notebook. A copy of Fulker's
draft IDD Principles document is also in the notebook.
Discussion highlights
- UPC estimates that the development time for "IDD in a Box" would be at
least two months and would require freeing Tom Yoksas from McIDAS support.
- UPC will control the topology for FOS and McIDAS data streams; WSI will
handle the distribution topology for NIDS.
- There was considerable discussion about how the UPC would handle any
disagreements over the implementation schedule.
- There was considerable discussion about the responsibilities of relay
sites and their role in the implementation. Domenico noted that the IDD is
a voluntary community effort; the UPC has no mechanism for enforcing
anything.
- The committee question the speed of implementation. The stated goal of
deploying 6-8 sites per month seemed slow to some, especially because it
reflects no improvements as installation experience is gained. The UPC
noted that experience showed that implementation problems are idiosyncratic.
Furthermore, experience also shows that distributing binaries doesn't speed
the process substantially. The problem is to get the right person and the
right machine and the right networking configuration all together at
implementation time.
- In establishing a topology, one goal is to reduce redundancy; while some
sites may implement alternate routes on their own, redundancy will cause
considerable network slowdown, which should be reason enough to follow UPC
plans.
- The goal of the IDD is to have no unsatisfied users. Realistically,
however, unhappy sites will undoubtedly appear.
- The implementation plan may be hardest on the "middle class": sites with
UNIX systems and a little experience cannot benefit from the the OS/2-only
solution and may not have adequate networking to qualify for early UPC
assistance (or IDD in a box) as relay nodes. In contrast, large
institutions probably can gain assistance by serving as relays, and the very
smallest may have all their needs met as OS/2-only sites or as points in the
topology where relaying is not needed (hence leaf node service is the only
practical alternative).
- There was some question whether the Policy Committee should pause and
reaffirm its commitment to the IDD or explore alternatives. It was agreed
that the risks associated with IDD remain acceptable and that the strategy
must be continued. Members reminded each other of earlier endeavors wherein
Unidata leapt into the unknown. The possibility of failure is a necessary
and acceptable risk in the Unidata environment.
- Brown commented that he had reluctantly come to the conclusion that the
community-based approach was the only financially feasible one to implement,
thus he supported it. He noted that it was also the approach most
consistent with the spirit of the Unidata Program. The IDD Principles
document is an attempt to articulate what this community effort entails; it
is not a legal document.
- In UPC publications and discussions, we should not condone the
widespread notion that networking is free to meteorologists. At the moment,
costs are borne out of an institution's overhead; this may not last.
Resolution 5:
The Policy Committee recommends that the UPC aggressively pursue connection
of Unidata members to the IDD as rapidly as possible. The community should
be informed as soon as practical of the UPC plan for accelerating IDD
connections to all Internet-connected members and of the information
requested from each member to enable connection. Priority will be given to
members willing and able to serve as relay nodes. All others will be queued
for connection on a first-come, first-served basis commensurate with
available resources and upstream capabilities after receipt of all
information needed for connection.
Action 6:
The UPC will develop queues for leaves and for relays and publish these on
the Web server.
Action 7:
The UPC will determine the current status of unconnected sites and Fox will
follow up on his letter 22 March 1994 to these sites.
Action 8:
Fox and the UPC will draft a letter to all sites informing them of key dates
(e.g., C-band termination date, sunset date of Unidata/Wisconsin stream via
satellite, etc.). The letter should encourage sites to work closely with
their regional networks (to keep abreast of impending network changes) and
with their campus network administrators.
Action 9:
The UPC will distribute IDD status reports (lifted from the UPC biweekly
reports) to the entire community.
Resolution 6:
The Policy Committee commends the UPC on their work on explaining and
clarifying the IDD concept and giving the Policy Committee a clear briefing
on this issue. The Policy Committee endorses the IDD principles document.
DIFAX
Fulker reported that Alden is willing to provide DIFAX over the IDD in
Group3 FAX form; the images would be compatible with the LDM. Fulker
believes that this would include the ability to create hard copies of the
images.
Discussion
- The UPC does not yet know how this fits with the Alden contract for FOS.
- There are other options: Unisys via FTP, broadcasts from other vendors
(e.g., WSI), Ku-band broadcast from Alden, Alden providing DIFAX via IDD
without UPC.
- Costs of the service are unknown.
- IDD in a box would probably accept DIFAX.
Resolution 7:
Given the current DIFAX situation and the high community interest in DIFAX
products, the Policy Committee rescinds resolution 9306.2 and advises the
UPC to address the cost-effective distribution of DIFAX products via the
Internet.
Unidata's Role in NSF
Jacobs noted that Unidata is currently living on a one-time transfer of
funds from the Geosciences Division in addition to its funding from ATM. If
Unidata is to continue receiving funds such as these, it will need to be
able to engage the interests of directors of other divisions.
Discussion
- Concrete examples of wider involvement are needed; adding seismic data
to the Unidata server at NSF was one example cited. Need to convert Program
Directors within NSF.
- It's important to show that Unidata's infrastructure works for a range
of sciences and with a range of data sets.
- As noted earlier, the strategic initiates will play a prominent role in
future budget considerations at NSF. By positioning Unidata to be a
Geosciences program, Unidata will be more eligible for both IITA and HPCC
monies.
Action 10:
Domenico will look for daily earthquake summary from IRIS and add it to the
NSF Web server if possible.
List of Resolution and Action Items
Resolution 1:
The Policy Committee thanks John Nielsen-Gammon for his many contributions
as a member of the committee and looks forward to his continued involvement
in the Unidata Program.
Resolution 2:
Given the report on Unidata/COMET workshop entitled Teaching Mesoscale
Meteorology in the Age of the Modernized Weather Service, the Policy
Committee recommends that the UPC review the implementation schedule of
netCDF compatibility and report on updated schedules at the February 1995
meeting.
Resolution 3:
The Policy Committee commends the Unidata Users Committee and all the
sponsors for the creation of a very successful workshop.
Resolution 4:
The Policy Committee will seek to broaden the Users Committee to encompass
representatives from the Geosciences.
Resolution 5:
The Policy Committee recommends that the UPC aggressively pursue connection
of Unidata members to the IDD as rapidly as possible. The community should
be informed as soon as practical of the UPC plan for accelerating IDD
connections to all Internet-connected members and of the information
requested from each member to enable connection. Priority will be given to
members willing and able to serve as relay nodes. All others will be queued
for connection on a first-come, first-served basis commensurate with
available resources and upstream capabilities after receipt of all
information needed for connection.
Resolution 6:
The Policy Committee commends the UPC on their work on explaining and
clarifying the IDD concept and giving the Policy Committee a clear briefing
on this issue. The Policy Committee endorses the IDD principles document.
Resolution 7:
Given the current DIFAX situation and the high community interest in DIFAX
products, the Policy Committee rescinds resolution 9306.2 and advises the
UPC to address the cost-effective distribution of DIFAX products via the
Internet.
Action 1:
Carry action items 1, 2, 5, and 8 from the March 1994 meeting over to the
October 1994 meeting.
Done.
Action 2:
Fox will issue a letter of invitation to Joe Friday and Dick Greenfield to
meet with the committee during the October PolCom meeting in Washington,
D.C. to discuss the status of international data commercialization and
exchange.
Done.
Action 3:
The Unidata charter may need to be stated in an independent document. The
statements from Chapter 5 of the Unidata proposal of May, 1992 to NSF were
distributed and reviewed. Fox and Bonner are to consider if and how to
clarify the Policy Committee charter and terms of reference, and will report
back at the next meeting.
Done.
Action 4:
The topic of Ynot will appear on the agenda for the October meeting.
Done.
Action 5:
The UPC will encourage the subcommittee preparing the workshop report to
redraft the resolutions so that the efforts involve the UOP/COMET and the
universities.
Done.
Action 6:
The UPC will develop queues for leaves and for relays and publish these on
the Web server.
Done.
Action 7:
The UPC will determine the current status of unconnected sites and Fox will
follow up on his letter 22 March 1994 to these sites.
Done.
Action 8:
Fox and the UPC will draft a letter to all sites informing them of key dates
(e.g., C-band termination date, sunset date of Unidata/Wisconsin stream via
satellite, etc.). The letter should encourage sites to work closely with
their regional networks (to keep abreast of impending network changes) and
with their campus network administrators.
Done.
Action 9:
The UPC will distribute IDD status reports (lifted from the UPC biweekly
reports) to the entire community.
Done. Placed on Web server.
Action 10:
Domenico will look for daily earthquake summary from IRIS and add it to the
NSF Web server if possible.
Done.
Index
Unidata Homepage
This page was Webified by Jennifer
Philion.
Questions or comments can be sent to
<support@unidata.ucar.edu>.
This page was updated on
.
