The University of Virginia and NSF held a workshop on
Scientific Database Management. A Summary has been published as
TR 90-21 and TR 90-22. Although it does not have a comparison
of actual file formats, it does make a useful guide for
evaluating various formats.
I would add that this is a subject in which it is hard to find "experts".
There are a LOT of domain specific "standard" formats. Some, like FITS,
THE standard in astronomy are well enough designed to be adapted for
general purpose use. Many are tied firmly to a specific problem domain
and/or file system structure. There are many ANSI and international
standards ( or proposed standards ) like ASN.1 and the proposed standard
interchangeable optical file systems that will have a big impact on any
"standard" way of representing data. ( not to mention new methods in
image/data compression like JPEG & MPEG standards, etc. )
This is a field that has fallen through the cracks for a long time:
MOST scientists have not been very concerned with HOW their data has been
stored ( until modern instrumentation has swamped them with data! ),
MOST computer scientists have not been concerned with such mundane subjects,
and MOST practical programmers have solved the problems on a ad hoc basis.
BTW: We (physiology dept. UVA) have been using HDF for multi-spectral
image files. We are considering using CDF for spectral (X-ray & EELS)
data. ALL of the standards we have looked at are insufficient in some
way and need to be extended. So I would say that one important criterion
is that there is a well planned facility ( both in the design and
architecture AND in the bureaucartic & administrative sense ) for extensions.
In that respect, the fact that there is a very active netCDF mail list
is a big plus.
======== "If you have a hammer, find a nail" - George Bush,'91 ========
Steven D. Majewski University of Virginia Physiology Dept.
sdm7g@xxxxxxxxxxxx Box 449 Health Sciences Center
Voice: (804)-982-0831 1600 Jefferson Park Avenue
FAX: (804)-982-1616 Charlottesville, VA 22908
