PARTICIPANTS: MEMBERS: VISITORS: Mohan Ramamurthy (Chair) Greg Byrd, COMET Greg Cox Brian Heckman, COMET David Dempsey Mary Marlino, COMET David Knight Bill Pennell, UCAR/UOP Steve Koch Jennie Moody Michael Morgan Charles Murphy Melanie Wetzel Doug Yarger EX-OFFICIO MEMBERS: UPC STAFF: Dan Vietor, ATAC Sally Bates John Merrill, Policy Committee Steve Chiswell Ethan Davis Ben Domenico David Fulker Linda Miller Don Murray Sandra Nilsson Mike Schmidt Tom Yoksas
A review of the action items led to a discussion by Dave Fulker of Action 3, pertaining to the use of IDD for distribution of regional models, e.g., MM5. Fulker reminded the UserComm of the principles document pertaining to the IDD. Fulker described the steps that Unidata is taking to make model data available to the university community using the LDM. Unidata is preparing a plan to be submitted to the NWS for a machine to be installed at the NWS/OSO for facilitating access and distribution to model data being made available from NCEP.
There were remarks by Committee members about data and metadata access which resulted in the following action item:
Fulker said that several advances in data management software should improve access to retrospective data throughout the Unidata community, a departure from the current Unidata model of focusing on real-time data access.
The area of display and analysis software will address the topic of advanced visualization software. The Unidata Users and ATAC committees will be tapped for their expertise and recommendations in this important area.
The development of Java tools will be done to provide a collection of platform independent packages for Java. Basic Java image viewers will be created to be used with McIDAS and GEMPAK, initially leading to a Java toolkit to provide more complex functions.
A receive only version of the LDM is being created in Java and should begin by mid-1998. This version will be similar to McLDM, except it will be available to the broader community. A full version of a high-speed LDM should be completed in time for NIDS (1999).
To provide for automatic IDD routing, Unidata is looking at multi-cast IP. It was suggested that Unidata check the telephone companies to see what they are doing, before spending a lot of resources in creating something that might be adopted from some other source.
There will be new products on NOAAport that will require decoder development, e.g. the GINI format. Within the GeoSif, a complete set of Java decoders will be produced by year 2001.
NetCDF development to address packing should be completed in 1998. A shift from NetCDF to Java libraries is planned. A client-server form will draw on the DODS work and should be underway by the year 2000.
The overall goal of the Geosif effort is to help the Unidata community utilize object-oriented methodologies and tools, especially the Java language.
Initially, a basic set of interactive meteorological applications with capabilities for storing and retrieving associated data will be created. Later, applications for analysis and display (including 3-D visualization), for handling data distributed among multiple locations will be completed for joining Unidata systems with Geographic Information Systems (GIS).
The UserComm recommended that Unidata track the existing visualization packages being used at universities to enable use of the data in a seamless manner.
Relating some of the timelines and data access items included in the proposal, Fulker reported that there will be a NOAAport Receiver at UCAR, hopefully by 4th quarter of 1998. A plan to have at least two universities purchase NOAAport receivers across the U.S., to provide redundancy and reliability was also mentioned.
EOSDIS should begin flowing through the IDD by the year 2001.
Fulker indicated that an additional three people are included in the resource requirements of the proposal.
During the NIDS discussion, a UserComm member said that some TV stations are using NIDS after the data is four hours old, rather than waiting for the 48 hrs in Unidata's contract with WSI.
Domenico reported that the University of Washington is using the LDM for distributing 88D Level II data.
The proposal is on line and available for review by the committee and was included in Section B of the notebook. Fulker suggested that the UserComm can provide input until Tuesday, March 18, 1997.
Fulker mentioned some early planning activities of a GPS network, with UCAR's UNAVCO and FSL, which is envisioned to move toward approximately 100 sites (including universities). The network will provide, among other measurements, Precipitable Water Vapor. It is envisioned that NSF-MRI funds could be available for cost sharing toward purchasing the GPS systems for universities.
Resolution 1: The Policy Committee recommends that the UPC develop an exit plan for OS/2 that results in the sunset of OS/2 by June 1998. This plan will be considered by the Policy Committee at its next meeting.
There are 50 sites using OS/2 on campus. This number reflects sites using a combination of Unidata's suite of software, McIDAS-X, GEMPAK and OS/2, including the OS/2 only sites.
Fulker pointed out that the context for the Resolution out of the Policy Committee is for the UPC to develop a plan for the future; one that would not depend on continued existence for OS/2. The plan should cover two major issues:
It was pointed out that the Resolution should not have been so specific and it should have provided for discussions across the community. The Resolution was seen as a potential for generating ill will. An alternative was recommended that a comprehensive approach was necessary and that additional information was needed.
John Merrill, acting in the capacity of Policy Committee liaison, said that the recommended plan is intended to provide a smooth transition for OS/2 users. Merrill indicated that it was the Policy Committee's intent to motivate an effort at the UPC to create a plan showing how the transition would work and provide for continued service to all Unidata participants.
The Unidata Users Committee discussed the OS/2 issues and suggested a counter resolution to the Unidata Policy Committee. It is:
Be it resolved that the Heads and Chairs encourage UCAR (Unidata, COMET, UOP), NSF, and AMS to build on the Unidata model to work with universities to:
It is envisioned that this project will have a strong university governance policy consisting of university participants; not a "rubber stamp" approach. The NSF has been very supportive of this initiative and wants to see it serve the broader geoscience and community colleges.
After several initial steps that have already been taken, a needs assessment will be done to find out what services should be provided. Investigation of multi-media that would be effective for the classroom, including tools, content, materials will be discussed. A proposal to the NSF will be created for funding. This money will not be taken from the research funds at NSF. Mary Marlino, previously from the USAF Academy, will lead the project within the UCAR Office of Programs.
The first year funding would be for development. During the first year, the following activities would take place:
This topic was raised due to a recommendation of a small group (6-7) of Unidata university users that Unidata adopt Linux as a Unidata-supported Unix platform.
The positive aspects of running Unidata software on Intel based platforms (ie PC's) were discussed. Several options exist for Unix based operating systems on PC hardware exist such as Linux and Solaris x86.
Given that the UPC supported packages GEMPAK and McIDAS are currently supported on other platforms that utilize byteswapped (Little Endian) architectures (eg Dec ALPHA and Ultrix systems for GEMPAK, OS/2 for PCs by McIDAS), the porting of these packages to a PC based Unix operating system has generally been possible within the current framework.
Unidata currently has one machine configured for testing Linux. A machine configured with Solaris x86 may also be possible in the near future.
At present, GEMPAK, McIDAS, and WXP have all been ported to Linux. The experiences of porting each of the packages were described to the committee. GEMPAK requires Motif which is distributed with Solaris x86, but must be purchased separately for Linux.
The support issues associated with PC based platforms were raised, including configuring hardware and drivers, as well as the mix of products available from different vendors. The general problem of obtaining support for Unix operating systems at sites was mentioned as a concern as well.
An action item encouraging Unidata to create an Intel-Unix mailing list which can be used by sites interested in discussing this platform/operating system combination was suggested. This would facilitate the exchange of information regarding the combinations of operating system and hardware configurations for which they have experience. The desire for this list to be indexed via the WWW in a manner similar to the way the other Unidata email lists are archived was expressed.
The UPC will create an email list for Linux users and track the community interest using Linux and Solaris x86.
Melanie led a discussion on the summer workshop framework. See Dave Dempsey's written summary (Section B in notebook).
The expectations of participants should be to learn how to teach remote sensing with scientific content and learn how to integrate the materials into the classroom.
Sandra Nilsson reported the following registered participants and breakdown to date:
Unidata has created a Web server location for creating materials to be included in the workshop and presentations to be made for the workshop.
There will be four groups assigned to the following subject matter areas:
Possible Software Needs:
Hardware Needs:
The projection system was discussed. This will need to be investigated further by checking with SCD to determine if their display system is the best solution for the needs of the workshop.
The next meeting, prior to the workshop, needs to be devoted to final coordination activities, the BAMS article and Exit Survey. The UserComm should attempt to collect information on future activities, such as regional workshops, meeting participant expectations and conducting a one year follow up activity.
There were three options discussed pertaining to bullet one:
Since option three is straight forward to implement, it was decided to conduct a test between Purdue and SUNYA. We are awaiting test results.
The second item (bullet two) was discussed among the Committee and it was felt that the Unidata community favors the higher resolution products over lower resolution products that have greater spatial coverage.
This link provides a the full list of the Actions and Resolution to the Unidata Policy Committee from the meeting.
Questions or comments can be sent to lmiller@unidata.ucar.edu
Go to the Unidata Homepage.