As I write this article, we are at the mid-point of the spring 1996 semester. How is the WIYN queue experiment doing? A quantitative analysis has been posted to the Web at http://www.noao.edu/wiyn/obsprog/. On those Web pages, progress reports are updated after every night allocated to the NOAO WIYN queue observing experiment. Please read those pages and send us any suggestions for improvement. We want them to be as useful as possible for you!
Qualitatively, spring 1996 is unfolding somewhat differently than fall 1995. For the most part, our feeling is that this evolution has been positive, but your comments would be welcome.
For example, during fall 1995, 70% of the usable hours (152/222) went to only three observing programs. This was driven by our goals to execute programs strictly in order of TAC grade and to complete programs at the 100% level whenever possible. This semester, we have relaxed these two rules. As a consequence, we are now servicing more programs with the goal of producing scientifically reasonable subsets of data for all "high priority" programs and as many "best effort" programs as possible. Of course, we will try to complete as many programs as possible. However, experience has shown most PIs prefer some reasonable amount of data over no data at all. We are carefully interacting with PIs to assure an accurate assessment of what constitutes a reasonable data subset when 100% completion is not feasible. This change in strategy has also eliminated the problem with programs greatly exceeding their TAC recommended time allocations.
Another positive development is the frequently updated queue activity information posted to the Web. Using this information (available at http://www.noao.edu/wiyn/obsprog/), it is now straightforward for PIs with programs in the queue to assess what is happening and when (or whether) their program is likely to be executed.
By reducing the number of programs initially entered into the queue, we have both reduced our program tracking workload and given PIs an earlier definitive answer about their program. Our goal was accept roughly an equal number of "high priority" and "best effort" hours, with the "high priority" hours being roughly 60% of the total available time. The idea was to give us a reasonable chance of servicing all "high priority" programs in the queue at the 80% or above level, and service as many "best effort" programs as possible. Unfortunately, as delineated on our Web page, we over-scheduled the spring 1996 more than we should have. In essence, this means that while most "high priority" programs will get significant datasets, many "best effort" programs will now receive no data at all. As we process proposals for fall 1996, we will keep this mistake in mind, and try to accurately schedule fall 1996 programs.
More information about the NOAO WIYN Queue Observing Experiment is available at our Web site. Please visit it and send us your suggestions for improvement.
Dave Silva