Re: Draft operational concept document

[mel wright 456 <wright@creek2.Berkeley.EDU>]

Hello Robert,

	nice outline. sorry I will not be at meeting.

A few comments

1. dynamic scheduling

We briefly tried an automatic dynamic scheduling program at Hat Creek
with the BIMA array two years ago. It used a prioritized list of
projects, and a simple decision tree including parameters like the
atmospheric path RMS as parameters to schedule the telescope. The code
can also be run off-line in a simulation mode. On-line, although it was
a technical success, it was a political and socialogical failure.
Observers really like being in control. The program is described in three
BIMA memos, available from the BIMA home page http://bima.astro.umd.edu

Hence my comment that dynamic scheduling should be instituted as the
default mode from the begining, and integrated with the proposal and
archiving.

2. on-line images and archiving images

Both the on-line images and the archived images used to access the data
need only be a small subset of the complete data.

The proposal writer should know what images, and/or other output will
be useful to produce on-line.
E.g.  images of integrated spectral line emission, or interesting
velocity intervals - not full 3D images.
If the spectra or integrated flux of a small region is of
primary interest, then this might also be an on-line output

There is other output too that is useful to access the
data quality - perhaps this should also be archived, along with
the images/etc which summarize the data.

Systems support astronomers could make sure that sufficient on-line
images etc. are produced to be able to decide if the observations
are doing what was intended.

3. archived uv-data.

This is the radio astronomy tradition, and is the primary data for
model fitting, so I suppose we should archive uv-data. Two points.

. Many calibrations are already done on-line. Many cannot be un-done. 
Archive calibrated uv-data with as many calibrations done on-line
as possible, and archive the calibration data if possible.

. no need to archive  the full correlator capability if only the
wideband or some subset of the spectra are of interest?  This may
preclude seredipitous discovery of an unknown spectral feature, but who
is going to look at all the data?

			Cheers,

				Melvyn.