Presented at the
"Jimboree"
held in honor of MIT Professor James Elliot on June 17, 2010
Finding out where spectra come from
It all started with Mars. Tom McCord, Jim's predecessor
at MIT took data on Mauna Kea with three different instruments
during its 1973 opposition. I got to reduce the data from the
newest of those, a 250 x 250 pixel vidicon spectrometer, which
covered the mineralogically interesting wavelength range
of 0.3 to 1.1 microns. The slit was in a 45-degree mirror,
and a photograph was taken at the time of the exposure.
To find out where we were looking on Mars, I projected
the photograph on a map of the disk of Mars at the time
of observation. I had to work out the map from first principles
as there were no widely available mapping programs at the time.
Due to the fact that the spectral features we were interested
in observing differences in across the disk of Mars were in
a wavelength range where there was lots of internal reflection
within the detector, we could never get the kind of data we
wanted, but the software lived on to be used with Mars spot
photometry which was observed through the same kind of 45-degree
mirror. For that project, I wrote software to reproject the
coordinates from the orthographic projection used for the
spectrograph slit and photometer apertures to a Mercator
projection which would show the entire planet. To write
that program, I had to delve into the MIT Map Room's library,
where I found only one book which actually gave the equations
I needed. The resulting coordinates went into my second
published paper in 1977 (even though the work was completed
in 1975).
Occultation predictions and finder charts
At first, we used predictions by other people, and I dusted
off my Mars mapping program to make a program called SKYMAP
which made finder charts, including scaled overlays for the
Palomar Sky Survey. I wrote my first catalog access software
to read the SAO catalog for these plots. Then we
Forays into Distances
The first contact I had with measuring extra-solar distances
started with occultations, too. With Deane Peterson, we used
lunar occultations to measure physical distances between binary
stars in star clusters, a rung on the galactic distance scale.
Spacelab 2 all-sky infrared maps
When I moved up the river to work on the soon-to-be-launched
Spacelab 2 Infrared Telescope, I took a tar tape of all of
the software I wrote on the MIT Unix system. Originally,
I was supposed to write mapping software in IRAF, but that
didn't work, so in a couple of weeks, I rewrote SKYMAP and
its subroutines to deal with all-sky projections for our
project. For a variety of reasons, the telescope didn't
work as planned, but the software and the K-band detector
did, so we got a map of the galaxy out of it.
RVSAO and Redshifts: Adding Z to X and Y
After the Spacelab 2 funding ran out, I put my IRAF knowledge to
work adapting several existing programs into an IRAF package
for computing redshifts by either cross-correlating object spectra
with template spectra or fitting emission line centers and
using their wavelength shift from rest. This package has been
widely used for redshift surveys of galaxies, getting the distance
to interesting extragalactic objects, and studying multiple stars
and exoplanets.
WCSTools: Catalogs and good coordinates for spectroscopy
In 1994, NRAO released an open-source library of astronomical
map projection software used by AIPS to put uniform world
coordinate system--pixel to sky direction mapping--information
into the headers of images. Astronomers wanted to match
optical data to radio data, but accurate world coordinate
systems were absent from most ground-based data. I started
out by writing a subroutine library for SAOimage to display
sky coordinates as an image is browsed. A descendent of
this library is now used by the widely-used ds9 image display
program. Then I used some existing software from the University
of Iowa and translated my image and catalog access software
from Fortran into C to make a package which could fit a world
coordinate system to an image by automatically matching point
sources in the image to stars in any existing source catalog.
Catalog comparisons
Over the next 10 years, I developed interfaces to each new
large and/or accurate start catalog as it was released. Eventually,
we wanted to know how good they really were, so I fit a set of
1736 wide field images taken by Warren Brown with world coordinate
systems and compared the residuals. New results for recent catalogs,
the USNO UCAC3 astrometric catalogr, the Sloan Digital Sky Survey
(SDSS), and the GSC 2.3, are presented here for the first time.
Redshift surveys with multi-object spectrographs
Recent activities, besides maintaining all of the software I've
talked about, include working on data reduction pipelines for
mutil-object and echelle spectrographs, used for spatial and
temporal surveys respectively.