The meeting was called to order at 6:30pm by Bob Bush. There were 21 attendees, representing 16 organizations, listed below.
The minutes of the December telecon were approved.
The meeting was adjourned at 8:15.
|Juan J. Alonso||Stanford University|
|Theresa Babrauckas||NASA Glenn Research Center|
|John Chawner||Pointwise, Inc.|
|Roger Davis||UTRC, Stanford|
|David Edwards||Intelligent Light|
|Steve L. Karman Jr.||Lockeed Martin|
|David Kenwright||NASA Ames|
|Steve Legensky||Intelligent Light|
|Mori Mani||Boeing St-Louis|
|Diane Poirier||ICEM CFD Engineering|
|Greg Power||Arnold AF Base|
|James Reuther||NASA Ames|
|Anthony L. Thornton||Lockeed Martin Skunk Works|
The CGNS Steering Committee has been established to help ensure the continuation of the CFD General Notation System (CGNS). In order to perform this mission most effectively, the Steering Committee would like to affiliate with an established organization that can support the mission of the Committee, and help ensure the establishment and continuation of the Standard.
The CGNS Steering Committee has become aware of the AIAA CFD Committee on Standards, and would like to explore the possibility of becoming a sub-committee of that organization. This affiliation will advance the AIAA mission and the CFD CoS mission to promote the acceptance of standards in a core area if interest to AIAA.
This document summarizes the expectations and anticipated working relationship that the CGNS Steering Committee would like to see materialize.
The AIAA Committee on Standards would create a new sub-committee, called the CGNS sub-committee, which would be the continuation of the existing CGNS Steering Committee.
The CGNS sub-committee would operate autonomously, and would not increase the workload on existing sub-committees.
A member of the CGNS sub-committee (usually the CGNS Steering Committee Chairperson) would represent the sub-committee at AIAA Committee on Standards meetings.
This affiliation does not preclude the CGNS Steering Committee from collaborating with other professional organizations that might also support the continuation of the CGNS Data Standard.
Access to AIAA communications media, technical conferences, committees, etc. will accelerate the process of gaining acceptance for CGNS and ensuring that CGNS meets the needs of the entire aerospace CFD community.
The AIAA would support the CGNS sub-committee by providing a legal entity through which contracts could be negotiated, and funding accumulated and distributed in support of the Standard. (The sub-committee is responsible for raising any funds that are to be disbursed by them.)
The CGNS sub-committee would abide by the CGNS Steering Committee Charter (until such time as the CoS adopts their own by-laws).
The CGNS Steering Committee has been established to help ensure the continuation of the CFD General Notation System (CGNS). In order to perform this mission, resources will be required to maintain and further develop the software and documentation, to help propagate the standard, and to ensure a sufficient support to new adepts of CGNS. This document outlines the effort the Steering Committee feels is required to maintain a viable data standard for CFD.
The effort has been broken down into the ten categories summarized in
the following table. An estimate in man-year of the level of effort
required in each given year is also shown. A brief description of the
specific tasks envisioned in each category is also provided.
|Tasks||Year 2000||Year 2001||Year 2002|
|Validation and Performance Reports||0.2||0.1||0.1|
|Platform/OS Upgrade Maintenance||0.2||0.2||0.1|
The more CGNS is being adopted, the more the users request extensions to the standard to support data types not yet part of the SIDS. The CGNS Steering Committee recognizes that it is not possible to satisfy every individual requirement. However, some requests are general enough to benefit the whole CFD community.
The role of the CGNS Committee would then be to determine which extensions are general enough to be added to the CGNS standard. It would supervise the elaboration of proposals by voluntary sub-committees and post these proposals on the CGNS web site for all users to comment on them. Once approved by the CGNS Steering Committee, these proposals would be implemented in the documentation and software (API).
As the CGNS API extends, and if more and more programmers add their contributions to it, the risk to introduce bugs increases. In addition, the more the user base grows, the more costly problems in the software will be. It is therefore important to establish a test metric that will grow with the API. This test metric should allow testing every function of the API and every node in the CGNS ADF tree. It should also allow testing both the C and FORTRAN interfaces on all supported platforms.
Users are questioning the performance of CGNS, and what they can expect in comparison to their existing I/O systems. Preliminary performance tests have shown that it takes more CPU time to create a new CGNS file, or to read an existing one, then to write or read a typical sequencial file. However, once the database is created, interrogating a CGNS file shows much better performance. This is due to the fact that CGNS not only writes long list of data, but also creates an organized database.
The penalty involved with the initial creation of the hierarchical database worries several users. In order to provide potential users with the confidence required to rely on the new standard, a thoroughly performance study on the CGNS I/O should be conducted. Parameters such as operating system, number of blocks and nodes in the mesh, unstructured or structured data, quantity of auxiliary information, etc, should be included in this performance study. The performance of the CGNS I/O in CPU time, wall clock time and memory, should be compared with other I/O systems (FORTRAN read/write, C fread/fwrite, Plot3D, etc... ).
The CGNS Data Standard (SIDS), and associated software (ADF and API), currently have a substantial body of documentation. This documentation will need to be enhanced, maintained and augmented, as extensions are approved and implemented. The existing documentation, although necessary and useful, is enough to scare away the most devoted CGNS fan... Since one of the main goals of the CGNS Steering committee is to propagate the standard by acquiring more and more users, it is necessary to provide for simpler and more comprehensive documentation, in the form of a User's Guide and Training Manual.
It is envisioned that as CGNS is accepted and adopted by various software organizations, that training will be required to facilitate new users. This effort will develop training materials, and provide training for new users. The aim is to reduce the barriers to full implementation, and to gain wide acceptance of the standard.
The CGNS web site was originally established as a mean to distribute the CGNS software and documentation. It has grown to be a key element in the propagation of the CGNS standard. In addition to its primary function of distribution, the CGNS web site also allows to recruit new users. It informs users of any pertinent CGNS news and receives user questions and requests. It is also used to post new proposals for extensions to the standard, and any other relevant documentation such as conference papers, presentation viewgraphs, etc. Finally it provides for a "meeting place" for all CGNS users. It can be foreseen that eventually a CGNS News Group could be added to the web site to facilitate exchange between users.
As the user base grows, the number of questions and requests send to the CGNS-Support Team increases accordingly. Timely response to users requests is mandatory in order to maintain users interest. So in the coming years, users support should be a priority. When the user base is large enough, we foresee that a News Group may be sufficient, as users could answer each other questions.
Utilities differ from extensions for they do not change the Standard Interface Data Structures (SIDS). They are instead tools facilitating the use of CGNS. For examples, it would be useful to have a utility to compare two CGNS files, or to view the file content graphically. Several proposals for utility development have been submitted to the CGNS Steering Committee. The Steering Committee supports the idea of creating some utilities that would benefit all users.
The CGNS API and ADF Core are distributed on most platforms/operating systems commonly used by the CFD community. Currently, these are SGI/IRIX, SUN/Solaris, HP/UX, IBM/AIX, DEC-Alpha/OSF, CRAY/Unicos and WindowsNT. Although not distributed on the CGNS web site, users have also compiled and ran CGNS successfully on Linux. It is important to continue porting CGNS to a wide range of platforms, and keep updating the compiled libraries as the different operating systems are upgraded. This automatically involves testing the API on all new or upgraded operating system.
In the early years of CGNS, the outreach effort is crucial in order to inform the CFD community of the existence of CGNS, and convince key organizations to adopt the CGNS standard, and maybe even contribute to its development.
Most administrative efforts are done on a voluntary basis. However some extraordinary administrative costs may arise. For example, the cost related to the user meeting at the Reno conference, or the cost of teleconferences between Steering Committee members. The Steering Committee should be able to cover for such special administrative costs.
Much of the work, to date and anticipated, will be performed as in-kind contributions from CGNS Steering Committee Members and CGNS users at large.
Additional resources will be sought to reimburse the software Focal Point for work performed to maintain the software.
Interested parties may seek resources to fund significant extensions to the standard or software utilities.
Funding/contributions will be sought from industry and government
agencies interested in extending and maintaining the CGNS
Standard. Support may be provided: