Summary of Target Selection Work Week, Jan 25-29, 1999

Goals: (1) Implement a pipeline which would select a scientifically
	interesting and reasonable set of targets from a stripe
	(runs 94+125) of data.

       (2) Identify people who would continue to code and debug
	modules, and monitor the selected targets.

	The code which has come out of this work week has met the goal
of selecting a scientifically interesting set of targets from a stripe
(runs 94 and 125) of data.  However, we have quite a lot of work to
do before we deliver a completed target selection pipeline.

	Here is a list of the participants with email addesses.  I would
like to thank each of them for major contributions to coding during
target selection work week.  These are the people who will continue to 
debug and monitor progress:

Xiaohui Fan 		fan@astro.princeton.edu
Steve Kent		skent@fnal.gov
Michael Strauss		strauss@astro.princeton.edu
Dave Schlegel		schlegel@astro.princeton.edu
Jon Loveday		loveday@oddjob.uchicago.edu
David Weinberg		dhw@payne.mps.ohio-state.edu
Wolfgang Voges		whv@mper5a.rosat.mpe-garching.mpg.de
Jim Annis		annis@fnal.gov
Scott Anderson		anderson@titan.astro.WASHINGTON.EDU
Jeff Pier		jrp@nofs.navy.mil
Jeff Munn		jam@nofs.navy.mil
Gordon Richards		richards@oddjob.uchicago.edu
Greg Hennessey		gsh@libra.usno.navy.mil
Scott Burles		scott@oddjob.uchicago.edu
Heidi Newberg		heidi@fnal.gov
Vijay Narayanan		vijay@astronomy.ohio-state.edu
Rich Kron		rich@oddjob.uchicago.edu
Ion-Alexis Yadigaroglu	ion@astro.columbia.edu
Jen Adelman		jen_a@fnal.gov

How priorities are set

	We hashed out several important design issues during work
week.  One of them was the method for setting priorities.  The previous
design had one four byte priority field which was overwritten by each
module which did target selection.  This meant that any object which
was targeted by more than one working group would only be considered
for targetting for the working group with higher priority.

	The new scheme sets the priority one of two ways.  If the
object was targetted by one of the tiled categories (hot standard,
quasar, or galaxy), then the priority is set by the tiled category
and cannot be targeted in a non-tiled category.  This is not a
problem, since it will be targetted if possible; there is no advantage
to trying again in a non-tiled category.  If the object was not
targetted by a tiled category, then the four byte field is divided
up into four one byte fields, with one priority in each.  The
target selection modules should return a number between 0 and 255,
where 0 means it was not selected as a target and 255 is the highest
priority.

	There is still some discussion about how the tiled priorities
should be set.  Since we try to get every tiled object, the
priority is only used to resolve conflicts between two objects of
different type which are within 55" of each other (the minimum fiber 
spacing).  However, we expect that there will be a significant number 
of such conflicts, so the priority scheme might affect the derived 
large scale structure, as the QSO correlation could be imprinted to 
some extent on the galaxy large scale structure.

Reports on science modules:

Galaxies
Quasars
Serendipity/ROSAT
Stars
Standard Stars