[fitsbits] FITS vis a vis XML

[Steve Allen]

We are building a fully-automated multi-object spectrograph for Keck,
and our original design specified that the observers would submit
a set of FITS HDUs with tables that describe the desired design
of the slitmasks.
In the original spec, the format for the HDUs of submitted data is the
same format which will be used for archival purposes in the FITS
tables which will be appended to the acquired image data.
Our specification includes keywords which describe the
interrelationships between fields in the different FITS tables in a
fashion that resembles the SQL-based relational database in which they
will be stored during slitmask fabrication and observation.
These interrelationship keywords are not standardized, which means
that they'll only be recognized by software cognizant of our
specification.

In recent ADASS meetings and online discussions it has appeared that
there is a growing contingent who promote XML as a panacea for all
sorts of data interchange problems.
It's pretty clear that XML provides a nice way of interchanging
tabular data between incompatible databases, and that there exist a
lot of whizzy and trendy tools for manipulating and viewing
XML-formatted datasets.
In FITS we already have a means of interchanging tabular data, and
thanks to folks at NOAO, STScI, HEASARC, etc.  we have some pretty
whizzy and trendy tools for manipulating and viewing them.
But the FITS tools are special-purpose tools known by a community that
is increasingly dwarfed by the XML tool community.

At the time of the project design review, going the FITS route seemed
like the right thing to do.
Having come to the stage of actually implementing the slitmask design
interchange we wonder whether the science user community might be
better served by using XML for submitting designs.
They will still have to use the FITS representation for the
archival data.

There is a great deal of material discussing the strengths and
weaknesses of XML with respect to various other formats, but FITS
is not one of them.  Is it time to contribute to a body of experience
of FITS vs. XML?

--
Steve Allen          UCO/Lick Observatory       Santa Cruz, CA 95064
sla at ucolick.org      Voice: +1 831 459 3046     http://www.ucolick.org/~sla
PGP: 1024/E46978C5   F6 78 D1 10 62 94 8F 2E    49 89 0E FE 26 B4 14 93