![]() |
![]() |
CCSM Software Engineering Working Group Meeting Summary
February 27, 2001 at Oak Ridge, TN
Focus: Software-Related Documents
The following is a summary
of the CCSM Software Engineering Working
Group Meeting (CCSM SEWG). It contains:
1) an overview of the purpose and results of the meeting; 2) a description of
some of the tools and guidelines that were mentioned, including URLs; 3) a list
of short-term action items.
1. Overview
URLs for the tools discussed at the meeting
are included later in this document. Currently software documents for the CCSM project are being
generated in a variety of formats (LaTeX, Word, html, DocBook), document
content is inconsistent (e.g., "design document" can mean very
different things), and auto-documentation tools are rarely used. Documents are
not that easy to generate, locate, or browse. The point of this SEWG meeting was to come up with a strategy to use
documents more effectively and generate them more efficiently. We proposed general guidelines for any
documents related to CCSM software. We
also proposed a set of basic documents and outlined their applicability and
content. The documents and guidelines
that were presented at the meeting are described at:
http://www.ccsm.ucar.edu/working_groups/Software/dev_guide/doc.03.html
The guidelines were accepted with the caveat that some
meeting participants wanted to try out the proposed LaTeX-based tools for
document generation before committing to a strategy. Comments on these tools or
on other aspects of the meeting can be sent to the ccsm-sewg@ucar.edu mailing
list or posted on the SEWG discussion forum:
http://www.ccsm.ucar.edu/cgi-bin/csm/BBS/bbs_forum.cgi?forum=software
We agreed to hold a "barn raising" before the
Sixth Annual CCSM Workshop in June, in which a handful of people will
instrument the atmospheric model using the ProTeX auto-documentation tool. If this goes well, there was interest in
instrumenting the other components in the CCSM as well. We will create and circulate templates and
examples for the documents proposed and for module headers. We may explore creating a position in NCAR's
Climate and Global Dynamics Division that would include assisting with
documents and maintaining the CCSM website.
There were some questions about what the priorities should
be in creating some of these software documents. Participants considered defining requirements and architecture
for the whole model and large model components important, since this helps to
communicate major design decisions. Documenting common utilities was also considered a priority.
2. Guidelines
and tools
One of the proposed guidelines is that documents be written
in LaTeX, since:
- it is a system many people are already familiar with,
- it produces reasonable print and web documentation using latex2html,
- there is a GUI interface option via the LyX tool, and
- it is compatible with ProTeX, a F90/C auto-documentation
tool.
Links to the latex2html, ProTeX, and LyX tools are included
here:
ProTeX: http://dao.gsfc.nasa.gov/software/protex/
(File List: protex
v2.0)
LyX: http://www.lyx.org/
latex2html
http://cbl.leeds.ac.uk/nikos/tex2html/doc/latex2html/latex2html
Some participants were curious about documentation produced
using latex2html. Here are some
examples:
NWChem: http://www.emsl.pnl.gov:2080/docs/nwchem/doc/user/index.html (thanks to David Bernholdt) MOM: http://www.gfdl.gov/~smg/MOM/MOM_manual.html
3. Action items
1) Try out and comment on tools (if interested, within next two weeks); 2)
Create templates and examples for proposed documents (DeLuca/Kauffman with CCSM-programmers, next two weeks); 3) Set
date for "barn raising" to instrument atmospheric model and coupler
(before June); 4) Explore adding additional staff to assist with and organize
documents, help to maintain website (ongoing).
4. Participants:
Jim Rosinski/NCAR
Brian Eaton/NCAR
Cecelia DeLuca/NCAR
Byron Boville/NCAR
Dave Williamson/NCAR
Tony Craig/NCAR
Tom Bettge/NCAR
Brian Kauffman/NCAR
Rob Jacobs/ANL
Jay Larson/ANL
Art
Mirin/LLNL
Doug
Rotman/LLNL
John Taylor/LLNL
David
Bernholdt/ORNL
John Drake/ORNL
David Erickson/ORNL
Pat Worley/ORNL
Bob Malone/LANL
Phil Jones/LANL
Chris Ding/LBL
Will Sawyer/NASA
ANL -
Argonne
National Laboratory
LANL
- Los Alamos
National Laboratory
ORNL -
Oak Ridge National Laboratory
LLNL -
Lawrence
Livermore National Laboratory
LBL -
Lawrence
Berkeley Laboratory
NCAR - National Center for Atmospheric Research
NASA - National Aeronautics and Space Administration