Policy Committee Meeting Summary
15-16 October 1998

Arlington, VA

Participants

Members	Representatives
John Merrill (Chair)	David Carlson (NCAR/ATD)
Ken Crawford	Harry Edmon (ATAC)
Robert Fox	David Fulker (UPC)
David Knight	Mary Glackin (NOAA/NWS)
Colleen Leary	Bernard Grant (NSF/ATM)
James Moore	Clifford Jacobs (NSF/ATM)
Charles Murphy	Jennie Moody (U.VA./Users)
	George Serafino (NASA/Goddard DAAC)
UPC Staff
Sally Bates
Ben Domenico
Joanne Graham
Jo Hansen
Linda Miller
Russ Rew

Administrative Matters

• The summary of the last meeting should be amended as follows:
  • Under the NOAA Report, the discussion point should read: "NOAA is funded through AWIPS build 4.2; future enhancements are being pursued."
  • Under NSF Report, the second sentence should read: "UCAR's funding for FY 98, however, was reduced by 1.6% ($0.95M), even though ..."
• The next meetings of the Policy Committee will be:

  28-29 January 1999, in Boulder, CO
  17-18 May 1999, in Boulder, CO

Status Reports

Director's Report and the Budget Report

Copies of Fulker's and Graham's slides were distributed at the meeting. In addition to his status report as given in the slides, Fulker noted that he'd made a presentation to the AMS Heads & Chairs on Unidata's Java transition and that it had been well-received. He also reported on receiving his medal from Educause/AMS .Graham gave the budget presentation, discussing FY98 year-end, FY99 planning, soft money status, and the resubmission of the 2003 proposal. She offered three scenarios for future funding to Unidata and discussed the implications of each.

• FY98 ended on a positive note with more than 200K in uncommitted carryover. However, it was noted that ongoing reserves were drastically depleted in FY98 as new funds to Unidata this last fiscal year fell short of the expenditure level by 300K . The expenditure level itself was not determined to be extraordinary, in fact was short 1.5 FTEs from previous years, causing concern for the future.
• FY99: steps have been taken to give the program the ability to track costs by both function and type. Planning for FY99 is underway and we taking a "spend slowly" approach until we get some clear guidance from NSF.
• Soft Money: Funding for non-NSF programs already in Unidata is stable, with no notable problems with either Case Studies or DODS funding. No new initiatives to bring funding to the program are in the works at this time.
• Worst Case Scenario: A scenario outlining the 5-year impact of budgets that are 3% above previous yearsÕ base identified a serious shortage of funds by FY00 year-end. In this scenario Unidata would have virtually no flexibility in times of change, and would, in fact, have to eliminate 500K in costs between FY00 and FY03.
• Medium Case Scenario: A better scenario in which Unidata receives a 6% increase in FY99 (above its 2.3M ongoing base), and 4.5% increases in subsequent years through FY03 was given. In this scenario, Unidata would run reserves to zero by FY03 and have very little flexibility, but would be able to operate without reductions in staff or current operating levels.
• Best Case Scenario: In this scenario Unidata receives a 6% increase in FY99 plus 300K. this new level of funding, 2.74M is then increased by 5% annually. If funded in this manner, Unidata would be able to function with some degree of flexibility and would be able to fund several advancement projects over the next five years.

Discussion:

• Alden's financial stability was questioned; it's in the throes of being bought by another company. Fulker noted that if Alden disappears, Unidata will still have three functioning downlinks and the MOU specifies that Unidata has the option to purchase the hardware at its depreciated value.
• Fulker was asked about staff morale in the face of budget stringencies, but he noted that this was difficult for him to assess.
• Graham's worst-case scenario assumes no year-end funds and a 3% increase for inflation in Unidata's base funding.
• Graham was asked whether Unidata could initiate a wage freeze; she responded that such an action was a UCAR-wide decision; Unidata cannot act unilaterally.
• Jacobs noted that he cannot increase Unidata's funding unless funding to NSF's ATM division is increased. He hopes that Unidata's budget will end up being somewhere between Graham's best and worst scenarios.
• The subcontract for Unidata's Web redesign is in response to the need to keep Unidata's baseline budget in line; it is not necessarily the best way to handle Unidata's Web work.

Users Committee Report

Copies of the Users Committee Meeting Summary are in the notebook. Moody noted that the meeting opened with Mohan Ramamurthy passing the Unidata chairman's bell to Moody. She then quickly summarized the meeting. The resolutions resulting from the meeting were on the Policy Committee agenda and there was no discussion.

NASA Report

George Serafino introduced himself, and commented on the noticeable lack of interactions between Unidata and NASA--a situation that Serafino hopes to remedy. He then explained that the Goddard Distributed Active Archive Center (DAAC) is customer-oriented. It is composed of three groups: customer support (Serafino's group), engineering support, and data management. The DAAC handles a range of data including such legacy sets as TOMS and Pathfinder, and some newer data sets (SeaWIFFS, TRMM); the DAAC will also archive AM-1 MODIS data (scheduled for launch in 1999), which may reach volumes of 650 TB/day. On new data sets, there is a two-day time-lag between observation and the release of the data (the same lag will apply to MODIS data). He noted that NASA has discovered that for most of its users, ease-of-use and ease-of-access are of paramount importance; thus, instead of relying strictly on nearline data access to large HDF files, they provide single-parameter flat files online that may be conveniently downloaded and interpreted using simple, DAAC-supplied readers that run on multiple machines.

Discussion:

• Serafino was questioned about the type of technology needed to archive 1,000 TB/day over six years. He reported that Goddard was installing an STK robot system and that it would evolve as needed. The MODIS data would grow to that size, but be much smaller initially.
• There was discussion about the possibility of archiving Unidata data streams, including the model data from NCEP. UPC staff noted that SSEC is already archiving nearly a month's worth of the McIDAS stream and also archives satellite data; Harry Edmon noted that he keeps roughly 7 days of raw data; and that individuals can always use the needdata list when they want some small, particular set of data; radar data is the most difficult to find. The SCD data park project was also mentioned. The problem, all agreed, is finding what you want when you want it. Domenico noted that this would be discussed later in the agenda.

Action 1: Serafino will check with George Huffman to see whether the Goddard DAAC group is archiving any of the NCEP data.

NOAA Report

Mary Glackin introduced herself and noted that Doug Sargeant (the committee's former NOAA representative) had passed on to her his shelf of information on Unidata. Glackin reported on four topics: NOAAport/FOS, WMO Resolution 40, NIDS, and N-AWIPS.

Concerning NOAAport, Glackin reported that the agency is engaged in an end-to-end test of the system's reliability and that it expects to announce at the annual AMS meeting that the system is operational. The agency is still adding FOS products, changing headers, adding gridded data (she expects NOAAport to carry more gridded data than FOS), and that current information on all of this is available on the Web.

Concerning WMO Resolution 40, Gen. Kelly has announced that, as of 31 October 1998, all international data received by NWS will be placed on the Web, but that a caution window will appear when a user selects these data, specifying that their use is restricted to education and that the data may not be redistributed. NWS will log the addresses of users accessing those pages. She also noted that there has been no official response from the countries most concerned about this action, only an "expression of dismay."

With respect to NIDS, the National Weather Service has established a requirement for the central collection of radar data to support numerical modeling efforts. Given a formal requirement, the NWS will be pursuing options to meet this requirement. The current NIDS arrangements are not acceptable for this purpose. Given this, the NWS has informed the NIDS vendors that the current agreements will not be renewed, but has requested to extend the current agreements 12-18 months to provide a transition to a new method of collecting and distributing NIDS data. The NWS is exploring several avenues for distribution: some of these data will be on NOAAport and some may be distributed using Internet push technology. The agency is required to send WFOs data from their local radars, but there will be no guarantee that these data will match exactly what the WFOs are currently receiving from the NIDS vendors.

Concerning N-AWIPS, Gen. Kelly has directed Glackin to specify an evolution of AWIPS to encompass N-AWIPS capabilities; a design team is currently examining what this might entail, but they do not intend to port the GEMPAK/N-AWIPS code to AWIPS. In the meantime, N-AWIPS is needed to interface with legacy software packages that are not Y2K compliant, so NWS will continue enhancing N-AWIPS for the coming year. By mid-1999, Glackin noted, N-AWIPS will still be in use at national centers, but it will no longer be in use at WFOs. Gen. Kelly is aware of the broader university concerns about the loss of N-AWIPS, but she noted that the commonalities between universities and WFOs have already been disrupted.

Resolution 1: The Policy Committee expresses its appreciation to Doug Sargeant (NOAA) for his long and highly constructive service to the Unidata Program and the community of Unidata users through his participation on the Policy Committee. (Passed unanimously.)

Discussion

• Universities need to be informed about the impending end to the NIDS contract; they should be advised to end their subscriptions by the end of 1999.
• The current NIDS contract covers NIDS products, not the Level II (raw) data. The NWS is examining the issues around providing access to raw data vs providing data that has some level of preprocessing.
• The UPC hopes that CRAFT, the U. Oklahoma prototype radar distribution network, will be successful; it could solve the problem of distributing Level II data. It is not an option for the NWS, however, since the agency is required to use dedicated lines, which would make distributing these data very expensive.
• COMET is being funded to document some of the AWIPS functionality.
• There were questions about the design team's approach to N-AWIPS. Is any code being ported? Glackin believed that the core N-AWIPS code would not be ported but the new N-MAP code might be.
• NIDS vendors are concerned about extending their contract because of the liabilities and efforts entailed by needing to be Y2K compliant.
• Toward the end of the meeting, the committee returned to a discussion of international data issues and WMO Resolution 40. The NWS has sent out a notice concerning restrictions to the use of international data. The UPC has forwarded the notice on the nws-changes mailing list. The notice is also available on the NWS and Unidata Web servers. The issue is moot for Unidata since the IDD is a controlled distribution system. The UPC can assist the NWS good faith effort to restrict distribution by splitting the international data into a distinct IDD data stream so that everyone in the Unidata community can tell exactly which data are involved.

Action 2: The UPC will send a letter to the community explaining the upcoming changes to the NWS NIDS contract; an article on this should also appear in the next issue of the newsletter.

Action 3: Glackin will determine what the plans are for distributing a national mosaic of radar data.

Action 4: The UPC will develop a plan for advising users on how to comply with NWS international data restrictions, identifying the costs involved, and report on this at the next Policy Committee meeting.

NSF Report

Copies of Cliff Jacobs' slides were distributed at the meeting. Jacobs focused primarily on KDI. He reported that the award process was sufficiently flawed that he could not recommend Unidata submitting another proposal to KDI. Bernard Grant reported on the distribution of equipment grants. He noted that the amount requested ($404,144) was unusually large because one proposal requested $100,000 in a series of tear-off proposals and another contained no cost-sharing provisions.

Discussion

• KDI declined to fund 11 proposals that were rated outstanding. Unidata's KDI proposal was not considered outstanding. Part of the problem was that the review panel members were drawn from such a broad range of backgrounds that achieving unanimity on proposals was extremely difficult.
• Cost-sharing is not required for the equipment grants, but it is preferred.
• NSF would need an additional $9,000 to fund all the equipment-grant proposals that the panel felt were worthy (11 of the 16 submitted); with the funding available, however, NSF could fund only 9.
• This was the first year NSF had announced the grants program through the normal NSF channels. As a result, there were more proposals than usual and one whose PI clearly didn't understand the Unidata program. Grant noted that he'd discouraged a few people from applying due to their unfamiliarity with Unidata and Bates noted that she had received (and discouraged) a few calls as well.

The Future of GEMPAK

At its September 1998 meeting, the Unidata Users committee expressed its appreciation for GEMPAK, particularly the ease-of-use brought to the software by the GARP interface. The committee expressed concern about the future of GEMPAK (N-AWIPS, N-MAP, and GARP) and asked that Fulker work with members of the committee to draft a letter to Gen. Kelly expressing these concerns. The software is widely used in the university community. The software also represents a point of commonality between universities and the WFOs. Fulker noted that the UPC staff would like to be able to employ GEMPAK ideas in the Java software now under development by Unidata.

Discussion

• The intellectual property rights to GEMPAK/N-AWIPS are unclear. Glackin remarked that NOAA is unlikely to pursue rights to N-AWIPS since they already provide the source code under the Freedom of Information Act.
• The UPC is still operating under the NSF agreement with NASA. (A copy of the agreement letter dated November 15, 1990 is in the notebook.)
• There was considerable discussion about how to obtain releases from NASA and NOAA. The UPC may ask NSF to get a release from NASA and NOAA via an MOU with no requirements to develop the software and allowing free and open access to the source code. There was concern about whether this was possible.
• There was a question of timing in pursuing a release of GEMPAK. There is little point in doing so while NWS continues to develop N-AWIPS.
• User concern about GEMPAK derives from fear that Unidata's only remaining easy-to-use package will disappear, as WXP disappeared.
• There was concern expressed that releasing the software to Unidata might result in "creating a monster" in terms of users expecting the UPC to continue developing GEMPAK. However, Unidata's proposal to NSF and the current budget crisis underlines Unidata's commitment to Java development as opposed to supporting/developing legacy software.
• There was some discussion about adapting AWIPS for university use. Glackin thought it was possible, but porting to a different platform would probably have marked performance costs.
• The message from the users is that Unidata should ensure that GEMPAK functionality is retained.

Action 5: Fulker will draft a letter to Gen. Kelly expressing an interest in being kept informed of N-AWIPS developments. The letter should indicate Unidata's awareness of NWS priorities.

Action 6: Fulker will look into obtaining a release of GARP from COMET.

Resolution 2: The Policy Committee recognizes the Users Committee's concerns regarding the future of GEMPAK/N-AWIPS and will make every effort to ensure that the functionality of this software is not lost to the community. The Policy Committee further recognizes the Users Committee's interest in maintaining some level of software commonality between the university and forecasting communities. (Passed with one abstention.)

Access to Retrospective Data

At its September meeting, the Users Committee passed the following resolution:

Users Committee Resolution 2 (passed unanimously):

In agreement with the recommendation by the NSF review panel, the Users Committee recommends that the Unidata Policy Committee explore Unidata collaboration with other agencies and organizations (e.g., NCAR SCD) to facilitate access to archived meteorological data in forms compatible with Unidata-supported software.

The Users Committee is particularly interested in having "easy" access to a data stored by SCD. Essentially, Moody explained, the Users Committee agreed with the NSF Review Panel's Recommendation 3:

Recommendation 3

The panel recommends that the UPC, in its development in general, work with UCAR entities and institutions in the development of tools and infrastructure that will facilitate access to remote historical archives. The panel feels that the two retrospective data sets of most interest to the university community are digital satellite and radar data. Additional data sets to consider are the daily climatic values from the National Climatic Data Center (NOAA) and the daily hydrologic values from the Water Resources Division (WRD) of the U.S. Geological Survey (USGS).

Moody noted that the NSF Reviewers' recommendation and the Users Committee resolution, when combined, identified a broad community need for access to archival data. Moody further noted that the committee recognized that this was not a task for the UPC.

Discussion

• There was concern expressed that the needs were too broad: the Users Committee needs to prioritize which data are wanted first. Since most Unidata users are meteorologists, the expectation would be that surface observations, upper-air data, and satellite images are of the highest priority among those in the SCD archive.
• Since there are established archives for all the basic meteorological data, the Policy Committee assumes that the Users Committee simply wants free access. But Moody pointed out that compatibility with Unidata software was also an issue.
• Unidata's needdata mailing list is able to solve some data problems, and there are lots of people archiving data in the community: Peter Neilly routinely archives FOS, so does Harry Edmon (a week's worth). Perhaps the need is for better communication within the Unidata community.
• UCAR does not have as its mission the provision of unlimited amounts of data, although each division tries to provide as much data as it can within the context of limited resources. One caveat, however, is that many of these efforts are ad hoc; making the archive known officially may result in putting individuals on the spot and requiring them to track costs.
• One major problem with the SCD archives is that a users must have an NSF grant in order to have access to the division's computers. Several people commented that there always is someone with an NSF grant somewhere on a campus, and this can be used.
• Rew explained that the Java tools under development at the UPC will have naming and directory services so the data formats of real-time data will no longer be an issue.
• Archiving of model-output data is a big problem; the UPC tools will not handle these.
• There are multiple data efforts underway by, for example, IRIS, UARC, DODS, CORE. These should be coordinated. The need for coordination should be incorporated into the GEO2000 vision document. Jacobs suggested producing a white paper delineating this need to gauge the reaction of the geosciences and to identify what a coordinated approach would entail. The white paper could point out that data formats represent barriers to interdisciplinary research.
• In addition to the white paper, creating a series of use-cases for archive data might help bring focus to the problem.
• The UPC has already identified access to retrospective data as an advancement project.
• Inviting an SCD representative to the next Policy Committee is an action item from the last Policy Committee meeting.

Action 7: Fulker will prepare for a more focused discussion at the next meeting; the topic will be data access and software compatibility , and an SCD representative will be present.

Action 8: Fulker will contact Dan Atkins (UARC), the NASA DAACs, and IRIS about concepts involved in facilitating access to retrospective data.

Action 9: Before the next meeting, the Policy Committee members will reflect on the relative priorities of accessing retrospective data vs other advancement projects.

Sense of the Committee:

The Policy Committee strongly supports Fulker's actions to bring a focus to discussions about access to retrospective data. A goal of this effort is to produce a white paper that can be used in discussions with other disciplines and agencies.

Unidata's Response to the NSF Panel's Reviews

Unidata's response to the NSF Panel Reviews had to be redrafted due to the NSF budget realities delineated at the last Policy Committee meeting. This agenda topic provides the committee with an opportunity to comment on the response.

Discussion

• There was some concern that several "near and dear" projects had been relegated to advancement projects (e.g., distributing Level II radar data).
• The list of advancement projects as presented to NSF were in priority order, but this was not specified in order to give NSF the greatest latitude in identifying funds for projects.
• Fulker was questioned whether the baseline would allow the UPC to keep up with technology advancements, particularly with advances in networking. Fulker noted that the UPC doesn't have to do this alone--all of UCAR has the same goal of keeping up. Fulker did express concern about the ability to enhance IDD. Dynamic routing, for example, will be needed if the radar data distribution system being tested at Oklahoma is successful and is to be widely implemented.
• Fulker was questioned about the UPC's flexibility in changing the baseline definition over five years. He responded that that budget is reviewed regularly by both the UPC and the Policy Committee and it is the role of Unidata's governance structure (namely, the Policy Committee) to alter priorities if needed.
• Moving the LDM into Java is not a baseline effort. As Java development proceeds, there will be increasing pressure to bring develop LDM in Java; this will have to be addressed at some point.

Platform Support Issues

At its last meeting, the Users Committee passed the following resolution:

Users Committee Resolution 1 (passed unanimously):

The Unidata Users Committee endorses the following changes in Unidata's platform support:
A. Immediately sunset support for ULTRIX and SunOS 4.x.
B. Add support for Red Hat Linux and Solaris for Intel (x86).
C. Sunset support for OS/2 on July 31, 1999.

Discussion

• To the UPC, sunsetting means ceasing development; it does not mean that the support staff refuses to answer questions about software or systems that are no longer supported.
• The UPC policy of giving long lead times before sunsetting a system doesn't need to be applied to ULTRIX and SunOS 4.x because no one in the community is using these.

Resolution 3: The Policy Committee adopts the following changes in Unidata's platform support as recommended by the Users Committee:

A. Immediately sunset support for ULTRIX and SunOS 4.x.
B. Add support for Red Hat Linux and Solaris for Intel (x86).
C. Sunset support for OS/2 on July 31, 1999.

(Passed unanimously)

Outstanding Participation Award

Sally Bates explained that the procedures for this award as passed by the Policy Committee results in few (if any) people being eligible for the award. Active participants remain active on committees and hence remain ineligible. In response to this impasse, the Users Committee passed the following resolution:

Users Committee Resolution 3 (passed unanimously):

The Users Committee requests the Policy Committee give the Users Committee full authority for making the Russell DeSouza Award. Current Users Committee members will be excluded from consideration. The Users Committee will seek nominations from the community, discuss the nominations in an open meeting, and render its decision in executive session.

Resolution 4: The Policy Committee gives full authority to the Users Committee for making the Russell DeSouza Award. (Passed unanimously.)

Unidata Membership Policy

Fulker noted that Unidata has participants, not true "members." The UPC does track software licenses and there are rules that restrict data usage, as required by Unidata's contracts with data vendors and the WMO Resolution 40 policy on the redistribution of additional data (see http://www.nws.noaa.gov/oso/addata.shtml). The issue of membership emerged in the UPC's negotiations with Alden concerning the distribution of NOAAport data. The MOU contains a de facto membership definition: "Unidata may extend the network to all current Unidata participants, to any degree-granting department at a U.S. or foreign institution, to any not-for-profit entity with UCAR, and to any government or not-for-profit organization with which Unidata is collaborating on matters of education, research, software development, or data access." There are no restrictions on who can use the McIDAS software or data streams; the UPC still licenses the software for tracking purposes. People wanting GEMPAK, but not falling within the listed criteria, historically have been referred to COSMIC (which is under transition). This is another reason for wanting to clear up the legal standing of GEMPAK.

Discussion

• The UPC was asked whether such an unrestricted membership policy causes support problems for the program. Fulker and Rew noted that there are both benefits and costs. Yes, more people need support, BUT more people contributed bug fixes, ideas for enhancements, and help in solving problems.
• There are more restrictive policies for receiving NIDS and lightning data, but defining and policing distribution is being handled directly by the vendors.

Communications Structures

Charlie Murphy explained that in the process of reviewing equipment grants, there arose questions as to how well Unidata was disseminating information about platform requirements. (A copy of Murphy's e-mail to Fulker on this topic is in the notebook). Sally Bates noted that this leads to the broader question of whether Unidata is communicating to the community effectively. She noted that the effects of the current restricted budget mostly affects communications structures, as can be seen from Documentation status report (in the notebook), which includes a list of items that have not been accomplished. Even though the UPC cannot accomplish all its former communications tasks, Unidata is exploring opportunities to augment communication based on sharing resources with other UCAR programs. Ben Domenico gave a presentation on what he's calling the Community Object-Oriented Hypermedia Library (COOHL), in which COMET, PAGE, and Unidata cooperate to develop a resources needed by all. One example is a tool for indexing useful resources currently on the web. (A draft explanation on this is available at http://fanny.unidata.ucar.edu/PAGE/CommunityAccess.htm. See also http://www.unidata.ucar.edu/projects/coohl/plans/commplan.html.) Domenico's presentation was followed by a free-form brainstorming session on communications.

Some Comments:

• You might allocate infinite resources to communications, but you still can't force anyone to access or read it.
• Many people are uncomfortable with e-mail based discussions. While one-on-one communications are the most effective, it is an enormous drain on resources and hence impractical in most instances.
• The Users Committee tried asking its members to contact people within the community, but few were diligent in doing so.
• There is often a communications gap within campuses. Most Unidata communications go to systems administrators, some of whom never pass the information on to others in the department.
• UPC might consider engaging the AMS Heads and Chairs in identifying mechanisms for improving communications. They are also a good resource for helping identify which sites COOHL should index.
• University administrators are always looking for resources that help their universities; Unidata might consider presenting itself as a resource.
• One on-line newsletter (Tropical Newsletter) is particularly effective in getting reader attention (a periodic e-mail of a table of contents; easy to unsubscribe.) Unidata might try emulating this.
• Unidata held forums at AMS and AGU, but we have not followed up on these.

Action 10: Bates will check with Julie Winkler on whether and how to engage the Heads and Chairs in an effort to improve communication.

Action 11: Bates will check out the Tropical Newsletter mentioned by Carlson and Fox.

Action 12: The Policy Committee will revisit this issue next year.

List of Resolutions and Action Items
October 98 Policy Committee Meeting

 

Resolution 1: The Policy Committee expresses its appreciation to Doug Sargeant (NOAA) for his long and highly constructive service to the Unidata Program through his participation on the Policy Committee. (Passed unanimously.)

Resolution 2: The Policy Committee recognizes the Users Committee's concerns regarding the future of GEMPAK/N-AWIPS and will make every effort to ensure that the functionality of this software is not lost to the community. The Policy Committee further recognizes the Users Committee's interest in maintaining some level of software commonality between the university and forecasting communities. (Passed with one abstention.)

Resolution 3: The Policy Committee adopts the following changes in Unidata's platform support as recommended by the Users Committee:

A. Immediately sunset support for ULTRIX and SunOS 4.x.
B. Add support for Red Hat Linux and Solaris for Intel (x86).
C. Sunset support for OS/2 on July 31, 1999.

(Passed unanimously)

Resolution 4: The Policy Committee gives full authority to the Users Committee for making the Russell DeSouza Award. (Passed unanimously.)

 

Action 1: Serafino will check with George Huffman to see whether the Goddard group is archiving any of the NCEP data.

Action 2: The UPC will send a letter to the community explaining the upcoming changes to the NWS NIDS contract; an article on this should also appear in the next issue of the newsletter.

Action 3: Glackin will determine what the plans are for distributing a national mosaic of radar data.

Action 4: The UPC will develop a plan for advising users on how to comply with NWS international data restrictions, identifying the costs involved, and report on this at the next Policy Committee meeting.

Action 5: Fulker will draft a letter to Gen. Kelly expressing an interest in being kept informed of N-AWIPS developments. The letter should indicate Unidata's awareness of NWS priorities.

Action 6: Fulker will look into obtaining a release of GARP from COMET.

Action 7: Fulker will prepare for a more focused discussion at the next meeting; the topic will be data access and software compatibility , and an SCD representative will be present.

Action 8: Fulker will contact Dan Atkins (UARC), the NASA DAACs, and IRIS about concepts involved in facilitating access to retrospective data.

Action 9: Before the next meeting, the Policy Committee members will reflect on the relative priorities of accessing retrospective data vs other advancement projects.

Action 10: Bates will check with Julie Winkler on whether and how to engage the Heads and Chairs in an effort to improve communication.

Action 11: Bates will check out the Tropical Newsletter mentioned by Carlson and Fox.

Action 12: The Policy Committee will revisit the communications issue next year.

Sense of the Committee:

The Policy Committee strongly supports Fulker's actions to bring a focus to discussions about access to retrospective data. A goal of this effort is to produce a white paper that can be used in discussions with other disciplines and agencies.