[fitsbits] FITS astrometry formats

Gentlemen and ladies,

I have the privilege of initiating the design of FITS format applications
for astrometry usage for ESO's VLTI, with particular application
to the dual beam PRIMA system.

I would appreciate pointers from those of you with astrometry experience
to examples that I might learn from, either radio or optical/IR.

A little background: PRIMA will be a dual beam, single baseline interferometer
operating mostly with the outrigger (1.8m) telescopes at Cerro Paranal at
K-band, with baselines to 200m. Our long term goal is to get relative
astrometry over fields of 10" or so to 10 microarcsec accuracy.

In addition to scanned fringes in three sub-bands of the K-band
we hope to record metrology data from the whole delay train. These
two data streams will be read at a few kHz. We will also record
engineering data from all of the active optical subsistems at a
similar rate, and environmental data at a slower rate. The recording
of each subsystem is generally asynchonous, but with a mountain-wide
master clock being repeated at each subsystem.

I want to store all this in FITS bintable formats, with the current IAU
standards (descended from the Jaffe/Cotton document) as a starting point.
In addition I want to store emphemerides, geodeitic information... that
has been used in the reduction.

Walter Jaffe

On Fri 2004-08-20T18:19:35 +0200, Walter Jaffe hath writ:
I want to store all this in FITS bintable formats, with the current IAU
standards (descended from the Jaffe/Cotton document) as a starting point.
In addition I want to store emphemerides, geodeitic information... that
has been used in the reduction.

Putting aside the use of pre-existing astronomical nomenclature, which
is certainly a good thing, this sounds like an application where there
could be many different tables with various interrelationships.

We have had good experience using the notions of relational databases
when designing the table formats. We have applied the normalization
notions found in database texts such as the one by Codd and Date. In
most cases we have tried to create table schemata which follow the
Third Normal Form. We find that the normalization is far preferable
to the kinds of empty fields or duplicated fields which tend to occur
in single "flat" table schemata.

Our tables have various primary and foreign key relationships with
each other. The entire content such FITS files can be ingested
directly into a database or extracted directly from a database by SQL
commands.


--
Steve Allen UCO/Lick Observatory Santa Cruz, CA 95064
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