Resolution 1:
The Policy Committee resolves that Policy Committee Resolution 96.1 be amended to read:
The Unidata Policy Committee resolves that the definition of those eligible to participate in Unidata include institutions of higher education within the United States and the World Meteorological Organization's Region 4. Other criteria currently in use for qualification of institutions to be participants remains unchanged.
[This resolution passed by a vote of 4:1.]

Resolution 2:
The Policy Committee acknowledges the need expressed by the Users Committee resolution on education as a statement of need by the community and passes the concerns expressed therein to Unidata Office of Programs director Bill Pennell for consideration by the ad hoc steering committee on education that Pennell convened in September 1996.
[This resolution passed by a vote of 3:2. A copy of the September 1996 letter from Pennell to D. Johnson, R. Johnson, and J. Snow re ad hoc steering committee on education is in the notebook.]

Resolution 3:
The Policy Committee endorses the Users Committee proposal for a summer 1997 workshop entitled: "Using Instructional Technologies and Satellite Data for College-level Education in the Atmospheric and Earth Sciences."
[This resolution passed unanimously.]

Resolution 4:
The Policy Committee notes and agrees with Users Committee Action 4, which states "Unidata needs to keep the community apprised through e-mail and Unidata Newsletter articles about the future directions being planned."

Action 3:
Include an article on proposal goals in the next Unidata newsletter.

Action 4:
The Policy Committee will consider the Users Committee Action 3 at a later date. (Users Committee Action 3 asks that Unidata explore the idea of using the IDD to distribute regional model data.)

ATAC Summary
A copy of the issues facing the ATAC is in the notebook.

Discussion

• SSEC is being funded by NASA to build its ADDE server for EOSDIS.
• Maintaining a reliable IDD system is the highest priority for the ATAC and for the UPC.
• There was concern expressed again that having relay sites on vBNS would result in a two-tier system; the UPC should begin urging sites to think about achieving higher bandwidth connections.

Action 5:
Distribute Otis Brown's UCAR Newsletter article to the community as soon as it is published.

NOAA Report
Linda Miller spoke with Allan Eustis, NWS, who said that the NWS had suffered a budget cut in the final appropriations for FY 1997. The results of the cut will delay AWIPS by four months and deploy 20 fewer units. A cooperative observer program to report snowfall measurements at over 200 airports is being implemented by the NWS. These observations will eventually be included in the FOS.

NASA Report
Fritz Hasler reported that money had been restored to NASA's budget for the GLOBE visualization project. It's based on an interactive spreadsheet model and is expected to be a tool for massive data sets. In creating the tools, NASA is focusing on the ability to display images. Hasler then demonstrated the tool.

NSF Report
Cliff Jacobs reported on the National Science Foundation budget; copies of Jacobs's slides were distributed at the meeting. Of particular note was the estimate of $3.33M of the FY 1997 NSF/ATM budget earmarked for ARI (Advanced Research Infrastructure) grants. Jacobs also noted that Unidata will be highlighted in the education chapter of the GEOSCIENCES long-range plan. Jacobs reported on efforts to establish the Open GIS Consortium and on the Methods and Models for Integrated Assessment (MMIA) effort. In Jacobs view, Unidata universities should be involved in both.

Discussion

• There was some discussion on what role Unidata might play in MMIA; it might be a source of funds, for example, for supporting the distribution of NIDS data. UCAR will be submitting a proposal.
• There was discussion of finding money for distributing regional model data and about the possibility of moving model output as a data stream.
• Jacobs did not spend any money on equipment grants last year so will have $200,000 for grants in this fiscal year.
• There was discussion about the possibility of applying for ARI funds to place NOAAport receivers at selected Unidata relay sites and for hooking sites to vBNS. Deadline for ARI proposals will the sometime in February or March 1997 so will need to act quickly

Action 6:
Establish a coordinating committee to look into an ARI proposal. Possible activities: NOAAport receivers at universities; relay equipment at universities for IDD; terabyte rotating archive equipment for archiving data. Nominations for committee membership would be to Fulker by October 18.

Action 7:
Fulker will talk to Pennell about establishing a UPC presence in the UCAR MMIA proposal.

Among the topics discussed:

• On the matrix presented by Fulker:
  • The activities appearing under "Wide Consensus" form a statement of work; the UPC is essentially promising to perform these activities--they represent the status quo.
  • The proposal needs to prioritize activities in the high-risk category. It was suggested that the matrix be redesigned into a pyramid with the activities subject to "first cut" appearing at the apex.
  • There need to be resource allocations associated with each activity.

• On writing the proposal:
  • Need to identify the benefits of new technologies (e.g., "the Model Ts are here, we now need to mothball the wagons").
  • There was concern expressed about that lack of involvement of vendors. In-kind contributions from such vendors as Sun, SGI, Microsoft, Netscape would immediately strengthen the proposal in NSF's eyes. Gaining development funds from vendors, however, was deemed impossible to achieve; better to try for one-shot donations of, e.g., equipment.
  • The process of writing a new proposal should include an examination and evaluation of Unidata's current partnerships; there should be a deliberate decision to continue or change these.
  • Many of the problems facing Unidata are being forced upon the community by technological change, not by the science. Need to stress:
    • process of decision-making;
    • range of potential routes (e.g., in the platform independence issue, the UPC could choose to support one operating system or move toward a network operating environment);
    • pitfalls of a particular course of action (e.g., the price of supporting legacy software).

• On the provision of visualization software (the proposed collaboration with SSEC's Hibbard on developing visualization software):
  • In choosing Hibbard's package, Fulker was employing the traditional Unidata paradigm (the software is freely available, its adoption is an example of working with a university, it represents a partnering in Java development); the selection was non-competitive.
  • Several members urged the UPC to look at other packages (e.g., IDL) with the notion of purchasing a "site license" that would apply to all Unidata sites; this would free Unidata from supporting the software. Otherwise, supporting a visualization package that will probably only be used by a small group (25-30) of sites will be very expensive. It was noted, however that VisAD and IDL are not functionally equivalent.
  • It was suggested that UPC's role in applications should be limited to scripting not to building from scratch.
  • There were questions about whether VisAD could be "sold" to the community; the ghost of YNOT was present.
  • Several members noted that having discipline-related functionality would be critical. The proposed collaboration on a Java implementation of VisAD begs the question of meteorological functionality ("UPC has identified how to build a tool box, but not how to use it").
  • Beginning in November, Vis5D will be distributed as part of McIDAS (and ultimately supported by the UPC).
  • Fulker had envisioned moving one support person to Java support.
  • Vis5D can read netCDF and HDF files.
  • "Free is not cheap."

• On Education:
  • May be an excellent source of additional funding for Unidata.
  • The need for infrastructure to promote NSF goals interdisciplinary/education/collaboration should be portrayed.
  • Unidata already has enough on its plate: the data issues alone are overwhelming.
  • Unidata's role should be clearly identified (and limited).
  • If Unidata's role can be defined as a technology "half-way house" for universities, then education technology might fit.
  • New data streams will require education on their use and applications.

• On Unidata's role in NOAAport:
  • There were questions about whether the UPC should be involved in running a NOAAport receiver. It was suggested that the UPC might ask Alden to become a partner: Unidata buys and owns the equipment, Alden runs it.
  • There should be more than one receiver in the Unidata community (a "diversity of sources").
  • The system should be one that runs without human intervention.

• On Unidata's relationship to small colleges and departments:
  • Some level of dissatisfaction with OS/2 exists.
  • Need to articulate the objective of using Java to make data access easier for OS/2-type sites.
  • Need to point out that the Unidata process will adjust the UPC's activities if it turns out that Java is not the route to the goal of achieving a network environment.
  • Web browsers have become THE GUI and universities have led the way in showing how the Web can be used.
  • Java will allow the packaging of data with the tools to deal with the data but still need to focus on the sources of data.
  • Historically, operating systems have been dropped/added as needed; UPC expects to continue these evaluations, particularly in the light of the needs of small departments.

• On the content of data streams:
  • Principles for evaluating which streams should be supported by the UPC are articulated in the principles for IDD participation.
  • Users Committee is the filter for deciding which data sets need to be supported.

• On commercial data management software:
  • Use of a relational database would help in indexing, comparisons, etc; it was suggested that the UPC consider purchasing a commercial package.
  • The UPC traditionally avoids requiring users to purchase software.
  • The UPC's use of CODIAC represents an example of Unidata building on the work of others.

• On the character of the Unidata Program:
  • The move to a network operating environment (Java) is essentially a change in Unidata's traditional role. Historically, Unidata has been a distributor (distributing data and software acquired/developed by others). Building Java-based tools moves the program into the role of builder. In this new model, how is the community involved?
  • Unidata is moving from adolescence into maturity. What are some of the implication of this? Should we consider changing the program's relationship to its sites? Urging sites to change their relationships to the public (i.e., charging for access to Web-based weather)?
  • Unidata has enabled sites to become "movers and shakers" in the world (the Daily Planet and Weather Underground sites were cited as examples.
  • What model should Unidata adopt for the future: an economic model? pay-per-view? Distributor? Builder?

Action 8:
Policy Committee members will ponder on the question: What is the Unidata Community?

List of Resolutions and Action Items

Resolution 2:
The Policy Committee acknowledges the need expressed by the Users Committee resolution on education as a statement of need by the community and passes the concerns expressed therein to Unidata Office of Programs director Bill Pennell for consideration by the ad hoc steering committee on education that Pennell convened in September 1996.
[This resolution passed by a vote of 3:2. A copy of the September 1996 letter from Pennell to D. Johnson, R. Johnson, and J. Snow re Ad Hoc Steering Committee on Education is in the notebook.]

Resolution 3:
The Policy Committee endorses the Users Committee proposal for a summer 1997 workshop entitled: "Using Instructional Technologies and Satellite Data for College-level Education in the Atmospheric and Earth Sciences."
[This resolution passed unanimously.]

Resolution 4:
The Policy Committee notes and agrees with Users Committee Action 4, which states "Unidata needs to keep the community apprised through e-mail and Unidata Newsletter articles about the future directions being planned."

Action 1:
Create an architectural diagram of Unidata software.

Action 2:
The UPC needs to set up a rolling budget for development efforts.

Action 3:
Include an article on proposal goals in the next Unidata newsletter.

Action 4:
The Policy Committee will consider the Users Committee Action 3 at a later date. (Users Committee Action 3 asks that Unidata explore the idea of using the IDD to distribute regional model data.)

Action 5:
Distribute Otis Brown's UCAR Newsletter article to the community as soon as it is published;

Action 6:
Establish a coordinating committee to look into an ARI proposal. Possible activities: NOAAport receivers at universities; relay equipment at universities for IDD; terabyte rotating archive equipment for archiving data. Nominations for committee membership would be to Fulker by October 18.

Action 7:
Fulker will talk to Pennell about establishing a UPC presence in the UCAR MMIA proposal.

Action 8:
Policy Committee members will ponder on the question: What is the Unidata Community?