Re: HDF, netCDF, etc question

[thompson(at)orpheus.nascom.nasa.gov (William Thompson)]

Aaron Birenboim <aaron@shimi.swcp.com> writes:

>Craig Markwardt <craigmnet@cow.physics.wisc.edu> wrote:

>: ronn kling <ronn@rlkling.com> writes:

>: * I advocated astronomy's FITS format.  Plus: platform independent,
>:   metadata is in ASCII, good support in IDL Astronomy Library.  Minus:
>:   seen as "archane."

>I agree.  Its very simple, very "sharable".

>: * Many people seem to swear by Liam Gumley's binary tools.  Plus: raw
>:   speed and direct access; platform neutral. Minus: low level.

>I dunno about this.  But before I advocated a simple, binary format
>tied to a database for metadata.  These could be a help here.

I also agree that a combination of a standardized format like FITS and a
separate database containing metadata is a good way to go.  For example, the
SOHO project uses FITS files, and is supported by an SQL database (Oracle)
which is populated by scanning the FITS headers.

We've also had good success using a database package written in IDL which does
not require a separate license for an SQL database.  This software, known as
the UIT database package, is available as part of the Astronomy User's Library
at

	http://idlastro.gsfc.nasa.gov/

At least one of the SOHO instrument teams uses this package to maintain its
catalog, instead of a commercial product like Oracle.  One of the major
advantages is that one can distribute the software and database files to
outside users without requiring them to buy anything other than IDL.

William Thompson