Policy Committee Meeting Summary: February 2-3, 1995

Boulder, Colorado

Participants

Members	Representatives	UPC Staff
Robert Fox (Chair)	William Bonner (UCAR)	Sally Bates
Russell DeSouza	Harriet Barker (UCAR)	Ben Domenico
Perry Samson	David Fulker (UPC)	Linda Miller
Carlyle Wash	Clifford Jacobs (NSF)	Sandra Nilsson
Julie Winkler	Douglas Sargeant (NOAA)

Administrative Matters

The summary of the November 1994 meeting requires the following corrections:
Meeting date: November 14-15
Page 8, bullet 4, should read "..., hence each campus client gets a $3000-6000 .."
(the hyphen was missing).
The next meetings of the Policy Committee will be
25-26 September, Monday and Tuesday, in Washington, D.C.
22-23 February, 1996, Thursday and Friday, in Boulder, Colorado

Status Reports

Director's Report
A copy of Dave's report is in the notebook; copies of his transparencies were distributed at the meeting.

Discussion

NCAR/UCAR now coping with Internet transition: trying to decide whether to tie in with a regional network or stay on its own. The vBNS network is to test applications requiring higher and higher speeds, unlike Unidata-type activities, which require a set bandwidth. Many NSFNET users wrongly assumed they would automatically become vBNS users and are only now beginning to cope with the impending changes.

Action 1:
If possible, Cliff Jacobs with arrange for another NSF networking update at the Policy Committee's Washington meeting in September.

There appears to be a tendency on the part of users to blame data interruptions on Alden even though the cause is often the underlying network. UPC expects this will change as users become increasingly familiar with how the IDD really works.
Fulker was questioned about the plans for netCDF. He noted that IDD development resulted in a hiatus in the development of netCDF, but the integration of netCDF with GEMPAK has been proceeding. The UPC anticipates netCDF development to begin again soon.
There was considerable discussion about the efficacy of having a Unidata booth at the AMS every two years, including concerns about the timing of the AMS, which precludes attendance by many students. (Unidata is represented in the UCAR Office of Program's booth on alternate years.) There was also interest in having Unidata reach out beyond the meteorological community. Creating a display for a new community of potential useers, however, could consume considerable staff resources.

Action 2:
The UPC will look into the cost of attending the AGU meeting in San Francisco and report back to the Policy Committee, either via e-mail or at the May meeting, depending on schedule requirements. As part of this investigation, the UPC will look into combining forces with UCAR, which traditionally has a booth at the AGU.

Budget Report
Copies of Sandy's transparencies were distributed at the meeting.

Discussion

UPC's spending rate for FY 95 is $2.6 M, but the program will get no more than $2.4 M next year.

User Committee Report
The Users Committee has not met since last October. Mohan Ramamurthy summarized recent committee activities as follows:

The report from the summer workshop has been submitted to BAMS. A report on the COMET short course was co-submitted.
Unidata users now have access to GOES 8 data through the research floater There has been some development of display capability in the community; meanwhile a subcommittee, headed by Melanie Wetzel is attempting to ascertain what capabilities the community wants. The subcommittee will present a proposal on this to the User Committee in April.
In response to discussions during the October meeting, the UPC has redesigned its Web server, adding a section devoted to details about various data streams. The staff is also working to get data providers to take up some of these tasks.
Users Committee members are now actively seeking community input by contacting sites directly.
The committee plans to survey users on software being used on campus. The committee wants to know what needs to be developed, supported, or made available to others.

Discussion

The committee was commended on its outreach efforts; Ramamurthy is concerned about the time involved, however. There was discussion about relegating this to sub-mailing lists, but felt that many people were still not using e-mail consistently.
Peter Neilly's WEATHER program in wide use in community; Neilly will be offering evening add-on session during next Unidata training workshop.

Action 3:
Users Committee will consider a mechanism for recognizing outstanding contributors to the Unidata Program

DeSouza noted that the GOES-8 floater was currently being used in support of a research project but Millersville intended to keep its center point on the East Coast with an 8-km resolution under normal operations. He also reported that two schools have signed up for the NIDS floater.

NASA Report
Fritz Hasler introduced himself to the committee. Copies of his transparencies were distributed at the meeting. He is Project Manager of the Public Use of Remote Sensing Data (RSD) Program, which was initiated by congress to provide access to NASA data.

Discussion

There is a tremendous potential for synergy between NASA and Unidata.

NSF Report
Cliff Jacobs reported that NSF received its budget earlier that week. Copies of his transparencies were distributed.

Discussion

Jacobs was asked about the possibility of rescissions; no one knows.

NOAA Report
Doug Sargeant, unable to travel to Boulder, gave the following report via a telephone conference call.

The current race to reduce government is having a demoralizing effect on NOAA. The goals for 1999 established during the last phase "reinventing government" are now expected to be achieved by 1996. This will significantly alter NOAA.
There are no major changes to report on AWIPS. The Pathfinder program is meeting schedules; a limited version of the NOAAport-like (point to multi-point) system will start in May or June (the data streams to be involved are known); NOAA continue to work with vendors on issues about access to NOAA data.
The future of lightning data is hazy; NOAA is considering a government-operated network.
A number of universities have requests to connect to individual WSR 88-D radar.
Sargeant encouraged Unidata's proposal to support SOOs; regards it as mutually beneficial. Also encouraged Unidata to continue its relations with FSL.

IDD Update
Ben Domenico summarized the recent milestones in IDD deployment. Details of the current IDD status are routinely available on the Web. His summary contained the following:

Milestones:
- LDM 4.1 distributed July 1994
- Field test with 30 sites continued into October
- Final release of LDM 4.1 was November 1, 1994
- 27 sites added to the system by December 1, 1994
- Mctingest and McLDM for OS/2 tested and deployed in 1994.
- By January 1995, Unidata had 65-70 sites participating in the IDD.
Measures of success:
- Everyone is on line that wants to be
- Sites seem pleased (comments are available on the Web)
- IDD delivers about 98% of all products every day
- In any given hour 70-80% of the reporting sites receiving 100% of the data.
Next tasks:
- Rearrange topology after the disappearance of NSFNET in May
- Advertise the December 1 cessation of the Unidata/Wisconsin KU-band broadcast.
- Bring in remaining sites
- Provide simple solution for new sites.
- Improve troubleshooting techniques
- Streamline how statistics are reported
Longer Term Tasks:
- LDM-5 with automatic fail-over (goal: late summer?)
- Provide more automated mechanisms for isolating faults
- Provide a simpler interface and more user training.
- Identify and use more relay nodes.
- Integrate with new technologies (e.g., multicast?)
- Continued adaptation of system to Internet changes.

Discussion

There was discussion of the copyright status of IDD. The LDM has the same copyright as netCDF: it's free to anyone and no one may profit from it.
When a site doesn't receive data, they can turn to sites acting as short-term archives.
Estimate that the IDD is moving gigabytes of data per day (180 MB of data times about 100 sites).

Platform Support Policies

Dave Fulker explained that there was a de facto policy operating within the UPC that resulted in cost/capabilities tradeoffs and allowed the UPC to respond to community needs and interests. He noted that while this policy resulted in staff porting software beyond platforms that were fully supported, this had the advantage of helping Unidata keep up with evolving industry standards. In practical terms, the Policy Committee should consider issues of legacy and the potential effects on and of the UPC environment. Essentially what platforms are officially supported are in large part determined by what applications are being used.

Discussion

Unidata represents a partnership between program and universities with the universities supposedly agreeing at minimum to upgrade their systems as needed. This can pose problems: universities are strapped for cash--can't afford every new OS; may have no control over some UPC "requirements" (e.g., Internet connectivity).
Shouldn't Policy Committee help identify what circumstance will yield a sunset? What will result in support for a new platform?
Within the UPC, have problems keeping OSs up to date and keeping software running on multiple versions of OSs. Have sunset older versions, e.g., IRIX.
Might consider establishing community support groups for some platforms.
Most of the current problems are ones of supporting different flavors of UNIX. This is not like a question of moving to an entirely new OS (e.g., NT). The bigger decisions should be considered by the Policy Committee on an ad hoc basis.

Resolution 1:
The Policy Committee recommends that the UPC, in partnership with the Unidata Community, continue to apply a pragmatic approach to decisions about which platforms to support and which to sunset.

Action 4:
Dave Fulker will articulate UPC's pragmatic approach to platform support.

Proposal to the NWS Regarding Support for SOOs

A copy of the proposal to NWS is in the notebook. Dave Fulker reported that he had attended an NWS meeting in which the NWS was struggling with questions of systems administration and software support. The meeting was so reminiscent of the Madison workshop that Fulker began considering how Unidata might help. The proposal was a result of this.

Discussion:

SOOs role is to do science and train forecasters. The NWS has bought platforms for SOOs that are to be used only for development, not by forecasters. This is a stopgap measured brought on by the slow development of AWIPS. SOOs have lots of new data but no tools for using them.
Everyone wants stronger NOAA/University ties
UPC would supply LDM software; NWS would arrange to acquire the applications software (GEMPAK is what's planned, but SSEC has not refused any request by any forecast office to use McIDAS to date.)
Opportunity for Unidata to reach a very different community with needs for support that are nearly identical to Unidata's. Represents a net gain for Unidata: back-up for Peggy Bruehl, access to case studies, and access to decoders and georeferencing.
There was some concern expressed that the proposal might represent an open door to support for other agencies. Agreed that these would have to be addressed as they arise.
Bob Fox expressed his interest in promoting the proposal as a mechanism for ensuring continued support by NWS of GEMPAK and as a mechanism for affecting the future directions of GEMPAK development.
Cooperation between NWS and universities already occurring (e.g., FSL and NPG); in fact there are about 25 such projects. The proposal would extend this interaction.
Proposal is designed to fit within existing MOU with COMET.

Resolution 2:
The Policy Committee supports the direction take by the UPC in regard to the UPC's proposal to the National Weather Service to provide support for SOOs, and asks the UPC to consult with the Committee if conditions (as outlined in the proposal) change.

User Committee Membership

Bob Fox would like to have nominations for members to the User Committee. Mohan Ramamurthy noted that nominees should be participants willing to give time, since the committee has a large work load. The Policy Committee then considered the nominations received to date.

Resolution 3:
The Policy Committee recommends that the Users Committee be expanded to nine members on a three-year rotation plan.

Associate Participation

The current need to defined policy in this arena is being driven by the Texas Natural Resources Bureau which would like access to data. The agency does not seem to understand that associate participation excludes access to data. There was concern that the agency would use data to make severe weather forecasts and UCAR's liability in such a case was questioned.

Action 5:
Associate Participation should be a major agenda topic for the next meeting.

Action 6:
The following sentence should be added to the cover letter sent in response to requests for associate participation:
"These systems are used for research and educational purposes and are not intended to be used in settings that require operational reliability."

List of Resolutions

Resolution 3:
The Policy Committee recommends that the Users Committee be expanded to nine members on a three-year rotation plan.

Status of Action Items

Action 1:
If possible, Cliff Jacobs with arrange for another NSF networking update at the Policy Committee's Washington meeting in September.

To do, closer to the meeting time.

Done. Nilsson will report.

Action 3:
Users Committee will consider a mechanism for recognizing outstanding contributors to the Unidata Program.

Done. Ramamurthy will report.

Action 4:
Dave Fulker will articulate UPC's pragmatic approach to platform support.

In progress.

Action 5:
Associate Participation should be a major agenda topic for the next meeting.

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 .