Policy Committee Meeting Summary:
May 22-23, 2000

Madison, Wisconsin

Participants

Members	UPC Staff
John Merrill (Chair) Michael Biggerstaff Robert Fox David Knight James Moore Charles Murphy Mohan Ramamurthy Julie Winkler Representatives David Carlson (UCAR/ATD) Allan Darling (NOAA/NWS) Harry Edmon (ATAC) Wayne Faas (NOAA/NESDIS) David Fulker (UPC) Bernard Grant (NSF/ATM) Clifford Jacobs (NSF/ATM) Mary Marlino (UCAR/PAGE) Dennis Thomson (UCAR Board) Tim Spangler (UCAR/COMET)	Sally Bates John Caron Joanne Graham Jo Hansen Linda Miller Don Murray Russ Rew Tom Yoksas Observers Thomas Achtor (SSEC) Steven Ackerman (SSEC) Thomas Whittaker (SSEC)

Administrative Matters

• The minutes of the January 27-28, 2000 meeting were approved with the correction of some typos.
• The committee welcomed new members Allan Darling, Wayne Faas, and Dennis Thomson.
• Action 1 from the January meeting has been delayed pending Kellie's schedule; Actions 2 and 4 were postponed pending the outcome of this meeting's strategic planning, and Action 3 was postponed due to the Users Committee's focus on the upcoming summer workshop, and Action 5 was postponed to the next Policy Committee meeting.
• The next meetings of the Policy Committee will be:
  October 5-6, Arlington, VA
  February 1-2, 2001, Boulder, CO

 

Status Reports

Director's Report
Fulker gave a brief overview of the program's status. He noted that Unidata staff had submitted five proposals to NSF (to ATM for the Users Workshop; to ITR with NCSA for work integrating netCDF on HDF; to ITR with U. Wisconsin and UC Santa Barbara for concept maps with VisAD, to EHR with PAGE/DLESE for core integration for NSDL, and to GEO with PAGE/DLESE for more work on the prototype Geosciences Digital Library). Fulker reported that the statistics on the IDD have been improved and are now more accurate, and they indicate that the IDD is stable. He also reported that discussion on how to provide Level 2 access to all 88D radars is continuing and that SuomiNet is near to purchasing 100 receivers, which UNAVCO expects to configure this summer. Finally, Fulker summarized that all of Unidata's core concern were coded "green" (in good shape) with the exception of the budget (coded yellow), which is still considered to be of concern beginning next year. Fulker noted that he expects to pursue new sources of funding in the near future.

Staff members gave brief status reports on their own recent efforts (full reports are in the notebook).

• On the LDM, Rew noted that he had identified the most significant performance and scalability problems with the LDM and has redesigned the LDM's product queue algorithms to eliminate them. He also expects to be able to increase the maximum size of the product queue from 2 gigabytes to a terabyte in an upcoming release. The feedtype limitations are being worked around until they, too, can be eliminated some time in the future. Anne Wilson is experimenting with NNTP as a new approach to data distribution.
• Yoksas reported that Unidata expects to release a new version of McIDAS at the end of July/beginning of August. (We received the new SSEC version the day after the Policy Committee meeting). Yoksas noted that two new activities are in progress: adding GOES sounder products produced by CIMSS to the Unidata-Wisconsin data stream using a new, more efficient data compression approach (PNG) and establishing a network of ADDE servers for the NOAAport channel's 1 & 2 GINI imagery. To date, the UPC and Plymouth State College (Koermer) are providing community access to NOAAport GINI imagery through UPC-developed ADDE secondary servers. It is hoped that at least two other community sites receiving the data may be willing to provide ADDE server access to the same imagery.
• Murray reported that MetApps will be developing new releases for demonstration at the summer workshop. Following the workshop, the developers will begin working to integrate the prototypes (e.g., integrating the 3-D visualization with the gridded data viewer).

Discussion

• There was discussion on the IDD statistics. The problems of a bottleneck at NOAA in the CONDUIT feed are skewing the IDD statistics unfavorably.
• Fulker was asked whether recent DOD action concerning GIS would increase the effectiveness of SuomiNet; Fulker believes it will not.
• The Unidata staff was questioned about the performance of Java; the staff has not experienced overall significant performance degradation in 2D, but volume rendering is slow in 3-D. Furthermore, some platforms (HP, for example) have no plans for supporting Java3D.
• There was discussion about the limited platforms able to run the MetApps prototypes. Unidata staff noted that this was the price for developing applications using cutting-edge technology. It was suggested that frustrated users might want to submit bug-reports to Sun about the Java limitations of Solaris.
• The staff was questioned whether Unidata could meet its proposed goals re MetApps by 2003. Staff members believe the program is still on track for this, particularly if a good component framework is developed.
• Jacobs noted that NSF benefits from knowing when its programs benefit industry. If bug-reports are proving useful to Sun, is there anyway to track this?
• There was some discussion about the importance of identifying Unidata's users. Unidata is developing a prototype tracking system and had planned to demonstrate its approach to the committee, but strategic planning has taken precedence.

Action 1:
Unidata will split out the statistics for CONDUIT in its IDD monitoring to enable a better view of general community data reception.

Action 2:
Unidata will publicize the minimum platforms requirement for running the MetApps prototypes.

Action 3:
At the January 2001 meeting, Unidata will report to the Policy Committee on its progress in developing a system to track its users.

Users Committee Report

The committee chair, Jennie Moody, was unable to attend this meeting, and she has resigned as the committee's chair. Charles Murphy summarized the last meeting, a full report on which is in the notebook. About 75 people will be attending this summer's workshop.

MetApps

Two people have been added to the task force. Eight members of the task force held a successful conference call and will hold more such calls in the future. The UMADA discussion environment is considered a clear benefit to most task force members. This summer, the task force will begin developing radar usecases.

Discussion

• The combination of ADDE servers and MetApps prototypes may provide an easy-access mechanism for groups (community colleges, etc.) that don't have sufficient systems administration resources to run an LDM.

Action 4:
A significant amount of time at the October 2000 meeting will be devoted to showing and discussing the MetApps prototypes.

NOAA Report
There were two new NOAA representatives at this meeting: Allan Darling and Wayne Faas. While neither one was asked to prepare a report, Darling presented the following:

• CONDUIT Status
  A hardware problem with the NWS router interface for the NWS/NASA T1 circuit was isolated and corrected in February. Data transfer rates returned to expected levels and have continued to be acceptable since that time. The NWS extends its thanks to George Huffman (NASA) and Steve Chiswell for their assistance in isolating the problem.
• NIDS Developments
  The NWS continues internal testing of data delivery via multicast and has begun testing multicast data delivery with external customers. Data for six of the 33 categories is now available via anonymous FTP in test mode. With the installation of an AWIPS modification, all data categories will become available on a site-by-site basis, beginning in late May. (See http://www.nws.noaa.gov/radfiles.html for a full list of data categories, FTP server data structure, and list of WSR-88D sites.)
  
  The NWS is beginning work to produce WSR-88D imagery for delivery on the Internet. GIF images for four data categories (19R Base Reflectivity [lowest elevation angle] 124 nm, 38CR Composite Reflectivity 240 NM, 80 STP Storm Total Precipitation, and 78/OHP One Hour Precipitation) will be generate in near-real time. Availability of these images is scheduled to coincide with the expiration of the NIDS contract 1 October 2000.
  • The NWS has produced estimates of data volumes for the WSR-88D transmission stream. For AWIPS v4.3 and v5.0 the typical volume (11 radars in VCP-31 mode) and the worst case volume (all radars in VCP-11 mode) are:
    • AWIPS v4.3 (Fielding now underway) 378,677 KB/hr 644,952 KB/hr
    • AWIPS 5.0 (Late `00) 569,830 KB/hr 996,020 KB/hr
  • NOAAport distribution of radar data in compressed form is scheduled to begin in the June/July time period. Data will still be encrypted, but periods of unencrypted transmission will be scheduled to permit testing. The Dissemination of WSR-88D Products Concept Paper and further information about the NOAAport distribution are available at http://www.nws.noaa.gov/noaaport/html/whats_nw.shtml.
• DIFAX
  Facsimile product distribution via DIFAX, which currently contains only FOS products, is scheduled to cease October 1.
• Decoder Development
  The NWS Integrated Work Team (IWT) is developing a web presentation to distribute decoder software. Details are available at http://www.nws.noaa.gov/tdl/iwt/. The team has released 8 of 23 scheduled decoder packages.

Discussion

• The NWS T1 being used by CONDUIT is a bottleneck (it's overloaded) and it is only funded for one year. The bandwidth of NGI is unknown today, but will have to be shared with other HPCC projects if funded. Darling was questioned whether FTP users would switch to NGI; he didn't believe so, since the NGI will not be set up as an operational service.
• There was discussion of the loss of products on CONDUIT. Since the LDM queuing is no longer time-limited, and since it's better to have all the products late rather than losing some products altogether (which makes the data useless), it was suggested that NCEP consider releasing the model output in stages rather than waiting for a model run to be complete before sending it out. Staging the data would serve to decongest the feed, and the committee felt this would be a tremendous service to the community.

NASA Report

The Policy Committee does not have a NASA representative at this time.

NSF Report

Cliff Jacobs presented an overview of NSF's budget for 2001 (an increase of 17.3% is being proposed). Several new initiatives are being proposed (IT, Nanoscale Science and Engineering, Biocomplexity and the Environment, and 21st Century Workforce). Programs within these may supply opportunities for Unidata (he mentioned particularly NEON and Earthscope) in managing data flows. GEO's portion of the budget would be about $40M for Biocomplexity, $16.6M for IT, and $1.5M for the Workforce. GEO has identified four focus areas: natural cycles, natural hazards, ocean observations, and global change. Jacobs noted that two reports will shape NSF in the near term: Shaping the Future: Geosciences Beyond 2000 (the printed summary of which was distributed at the meeting) and the National Science Board's report on Environmental Science and Engineering for the 21st Century. The NSB recommends that NSF receive a $1B increase in funding and become highly visible, that it's projects should be considered a long-term investment, and that enhancing environmental education be a priority. (NSF is now required to assess the scientific value of its environmental programs.) Finally, Jacobs noted that sustainability is becoming an important focus for NSF.

Discussion

• Jacobs observed that in the past individuals have played critical roles in the advancement of science (citing Walter Orr Roberts role in founding UCAR and Packard's role in promoting ocean science).
• Jacobs was questioned about NSF plans to fund an NCAR-like facility for biocomplexity or sustainability. Jacobs said there were no such plans. He noted that part of the difficulty is that few sciences match atmospheric science in their level of collaboration, sense of community, or global perspective.
  

PAGE/DLESE Report

Mary Marlino gave a brief update on PAGE/DLESE and its relation to Unidata. She noted that she and Fulker recognized that their missions are related and that they share core values of 1) community governance/ownership; 2) users as contributors and as active participants, and 3) distributed responsibilities: the digital library will be a distributed, federated facility (although she noted that no one knows yet how to do this). In PAGE's Geoscience Digital Library (GDL) effort, the program is building a prototype library. DLESE has a steering committee, a community plan, and is about to publish articles of federation. Marlino noted that digital libraries to date have failed because of a lack of funding and a lack of community involvement.

Discussion:

• Marlino was asked how DLESE would be different from, e.g., EOSDIS? She noted that DLESE would be geared to a different audience, one of educators who want enriched collection. DLESE will be more than a catalog or a repository of data; it should be a community center.
• Assessment is part of the effort; there will be several levels of collections.

Strategic Planning

A draft ("strawman") strategic plan is in the notebook, and a revised edition was distributed at the meeting. Fulker noted that this planning session should focus on ideas for the long term (5-10 years) in an effort to keep the Unidata Program flexible and as a hedge against being "overtaken by events." He pointed out that the draft reflected discussions among Unidata staff members, and assumed that Unidata's constituency would remain academic, and that its mission would continue to include a focus on building community; other assumptions underlying the draft are in sections 5 and 6.

Participants divided into four groups to address the mission and vision statements in the draft plan. The proposed mission statement was "Providing infrastructure and software for use in academia to observe, understand, and share knowledge about Earth systems." The vision statement is in the revised draft. A summary of the groups' discussions are as follows:

Group Leader: Charles Murphy

On the mission statement:

• The word "observe" is confusing; it brings instrumentation to mind rather than data manipulation or distribution.
• "for use" is too weak and "providing" implies a one-way relationship between the program and its users. The statement should be proactive and include the idea of enabling community (two-way and multiple interactions); " building" was suggested.
• Software implies products and is subsumed by infrastructure and is therefore unnecessary.
• Can't make community-building a focus; it's an outcome; need to support collaboration.
• Sometimes the most effective path to a goal is not a straight line (e.g., airlines' and Fed. Express use of hubs). Unidata's involvement of users may seem inefficient but is effective. Unidata's role is to pick standards
• The expansion to Earth system is an enormous change, but it's necessary.

On the vision statements:

• Globalization is happening and the international flavor of earth system needs to be acknowledged.
• Unidata needs to provide not just "easy access" but "easy and equitable" access, which is hard to do.
• There was concern about defining the Unidata community: it'll become larger, more diverse, and harder to track; communication is going to be a big issue. Involvement with DLESE will be instructive since it has a less well-developed community.
• Unidata's leadership role may be less active, more focused on picking the "right horse" (standards and technologies).

Discussion:

• Who is the strategic plan being written for? The question was unanswered.
• What is meant by community-building? One remark was that fostering community is a function of the funding agencies.

On the goals:

• The group found discussing this more difficult. Questioned current work vs. future work; the goals seemed directed toward current work.
• Goal 1: is a current goal and Unidata is just now working toward the "self-management" part.
• Goal 2: have packages, now are working on Java tools
• Goals 3-5: seem new to Unidata, but in #4 have UMADA working now.
  • these goals are vital to Unidata's new mission
  • in #4, in terms of distance education, current tools are ineffective; also, collaboration and communication may be separate goals.
  • addressing problems of data formats and geolocation is needed for effective collaboration
  • the future is getting more complex; need to keep things as simple as possible.
  • in #5, need to engage new users, need to advertise current collaborations; need to have mechanisms for informing others of new efforts, which means learning more about them to identify common problems.

Discussion:

• There was considerable discussion on whether hiding complexity was a good goal. Some felt that hiding complexity generally means losing flexibility. Others noted that complexity is inescapable: expanding to ESS means more complexity, a diverse audience means more complexity, the advances of sciences means more complexity.
• The community needs to be able to see all of Unidata's goals; complete transparency would not be a good thing.

Group Leader: David Knight

On the mission statement:

• "Earth system" is too broad and is unrealistic; however, no alternative was suggested.
• "use in academia" -- the restriction is a good idea, but the group preferred using "education and research."
• "Providing infrastructure" implies that Unidata will providing networks to universities.

On the vision statement:

• There was little debate; no holes were identified, but the vision statement needs to be shorter and more powerful.
• Group suggested the following changes/comments:
  • Bullet 1: is an assumption; move to that section
  • Bullet 2: needs wordsmithing ("cross-disciplinary display software")
  • Bullet 3: is an assumption; move to that section
  • Bullet 4: needs wordsmithing: suggest shortening to "powerful visualizations used with multifaceted data"
  • Bullet 5: an important PAGE/DLESE activity
  • Bullet 6: needs to be shortened
  • Bullet 7: is an assumption; move to that section
  • Bullet 8 & 9: combine in vision of new forms of collaboration
  • Bullet 10: delete
• The "immense transformation" is unclear: is it Unidata's or the community's?

Discussion:

• Bullet 10 might be reworded to introduce the international component. As written, it smacks of political correctness, yet problems of diversity must be addressed, maybe in the assumptions?

On the goals:

• Goals are good.
• The discussion was remarkably peaceful
• Group had only minor suggestions:
  • Goal 1: change to availability of "real-time data"
  • Goal 3: need to specify retrospective data and educational materials
  • Goal 4: delete "diverse tool"
  • Goal 4.2, first bullet: "continuity AND EXPANSION of ..."

Group Leader: Julie Winkler

On the mission statement:

• The character/purpose of a mission statement:
  • something that can be memorized (short!)
  • is not limiting but does provide guidance for setting priorities (i.e., is not restrictive)
  • describes things that could be done
  • is something that will be around for awhile
• Infrastructure is a "loaded" term; "observe" could be misinterpreted.
• Suggested alternatives:
  • Short: Unidata: A catalyst for Earth systems research and education
  • Long: Enabling academia to understand and share knowledge of Earth systems by providing technology and community building.
• Discussed the term "Earth system":
  • What are the implications for Unidata (e.g., a change in Policy Committee composition?)
  • Is Earth system accepted to mean earth-atmosphere-ocean?
  • What about space science?

On the vision statement:

• Suggest reorganizing this into two parts:
• Part 1: Vision of the Community (Bullets 1, 4-10)
• Part 2: Vision for Unidata (Bullet 2)
• Bullet 3: delete
• Bullet 4: add use of local data and local models
• Bullet 5: invert sentence
• Bullet 7: this is "loaded" and is the only bullet with no highlighting.
• Bullet 10: delete "and reflects a broader range of ... models."
• The first paragraph (along with first sentence from the last paragraph) will be the heart of the vision statement.
• Some wordsmithing:
  "In this immense transformation, Unidata will play a catalytic role by providing technologies for acquiring, organizing, and using geoscience data and for the fruitful discourse on Earth system education and research. As in the past, community members will be engaged as owners of and contributors to the future."
• The statement is missing the advocacy role of behalf of the university community to NOAA, NASA, etc.

On the goals:

• Goals weren't exciting; seemed to be for existing work, which is acceptable since the evolution will be incremental.
• Missing community-building; add to 4.4 or 4.5
• A more logical progression would be 4.1, 4.3, 4.2, 4.4, 4.5
• Goal 4.1: Unidata's bread and butter; local data needs might be separate goal
• Goal 4.3: Separate goal to enable data retrieval for low-bandwidth sites
• Goal 4.2: Make it clearer that Java tools are improved tools that enhance capability; add the porting of displays into presentation software. Also missing interactions with GIS
• Goal 4.3: List last two objectives in separate goals.
• Goal 4.5: explicitly include off-site training. Also must deal with two separate diverse groups (non-atmospheric disciplines and small institutions).

Discussion:

• There was discussion on what constitutes local data (e.g., agricultural station data? other national observing networks?)
• Need to address data transparency and data performance
• Need to make it easy for people to make their local data available to others

Group Leader: Mohan Ramamurthy

On the mission statement:

• Want more emphasis on leadership and a proactive role for Unidata ("enable" is too passive)
• Prefer "researchers and educators" over "academics"
• Agree with broadening to the "Earth-science" community, as long as it's done incrementally with attention to avoiding disruption of the existing core community.
• Emphasis on "enabling good science" as the ultimate mission.

On the vision statement:

• Existing bullets should be moved to goals/tactics/assumptions
• Use instead:
  • deliver information in a form most easily used by educators and researchers for the advancement of science.
  • make Unidata infrastructure become such an integral part of the community that it becomes transparent to the users.
  • continue to exploit and influence emerging and existing technology for the benefit of the scientific community.

Discussion

• If you wanted to identify what Unidata would look like, it would be partially invisible.
• CISE and ESS are beyond Unidata's core community; Fulker may want to get input from them.

On the Goals:

• Goal 1 is a subset of 3 and is missing community-building goal (advocacy, services, leadership, credibility). These aren't covered in goal 4.
• Goal 2 is too narrow: visualization plus metadata and complementary materials to enable knowledge discover (e.g., uses of GIS data)
• Goal 5: delete the world "tool"; the group was divided on the interpretation of "diversity"; needs clarification

General Discussion

• Growth will need to be incremental and managing the core community will be a delicate problem.
• Expanding to community colleges is an implementation issue; it may need to be subcontracted. Jacobs believes this expansion would not be embraced by Unidata's current community.
• Perhaps it would be best to view Unidata as "enabling but not requiring" expansion to include community colleges through advocacy and by creating attractive technology. However, would still need to address the issue of engaging the expanded community in participation.
• Expansion into international arena has the added problem of data issues. There was some discussion of excluding users from nations with restricted data policies, but others felt that this only penalized users who are already victims of their governments' policies. It was also noted that Resolution 40 doesn't apply to retrospective data. Involving international users would help gain a true global perspective.
• Researchers and educators have different goals. Educators may want to integrate a greater variety of data and thus may be easier to engage.
• Users need data on demand irrespective of time.
• Unidata has not been effective in providing access to archival data. There were questions about the role of DODS, which is primarily a data server. Users will need clients and sites to host the servers in order for this effort to scale. Fulker noted that this is being addressed by SCD's director as well; both groups needs to define a vision for how to serve retrospective data. There was also the question of where Unidata will sit in respect to EOS.
• There was considerable discussion about the meaning of diversity. Unidata is trying to insulate users from the complexity of underlying technology but should not be assuming the role of explaining the content of material accessed by that technology. However, Fulker noted that DLESE will be assuming some of the "explaining" role and that accessing contextual information is closely aligned with technology issues. Furthermore, it was noted that NSF now recognizes the problem of enhancing scientific understanding and sees Unidata as "serving society" when others can apply tools in a variety of context.

SUMMARY

Fulker summarized the strategic planning discussions as follows:

• New mission statement may be something like "A catalyst for Earth system science education and research."
• He will add a leadership role in developing technology for acquiring, using, and analyzing ESS data and for enabling collaboration.
• Need to emphasize maintaining Unidata's core community
• Need to add notion of community-building.

He noted that Unidata is facing some hard questions:

• How to embrace a broader community
• How to address issues of diversity
• How to become interdisciplinary
  • perhaps use the "bait" of data flows and servers
  • through GIS systems
  • through cooperation with DLESE
  • through MetApps, but developing general tools
  • by integrating tools and data in educational materials.

He foresees the discussion continuing via e-mail or by other mechanisms.