WIYN Queue Observing Experiment: Fall 1995 Summary (1Jun96)

(from KPNO, NOAO Newsletter No. 46, June 1996)

In the last NOAO Newsletter (No. 45), the lessons learned from the fall 1995 semester of the WIYN Queue Experiment were discussed in general terms. With the completion of fall 1995, a quantitative summary can now be presented.

On 1 February 1996, all fall 1995 WIYN queue programs were officially terminated. PIs were informed that to continue their program, to start a new program, or re-start their old program, they had to re-apply for observing time.

As discussed in NOAO Newsletter No. 45, the WIYN queue was heavily oversubscribed in fall 1995. Consider the following tables:

	WIYN Fall 1995 Subscription Statistics

      Number of Proposals           Number of Nights 
	Submit   Sched   Ratio   Request   Avail   Ratio 
Dark      36      26      1.4      80       34      2.4 
Bright     9       6      1.5      27       16      1.7
2hrQ      20      16      1.3       5        4      1.3

	WIYN Spring 1996 Subscription Statistics

      Number of Proposals           Number of Nights 
	Submit   Sched   Ratio   Request   Avail   Ratio 
Dark      47      16      2.9      122      36      3.4 
Bright    17      10      1.7       52      13      4.0
2hrQ      22      13      1.7        5       3      1.7

If one considers the fall 1995 "Nights Requested/Nights Available" ratios, you can see that the dark-time and bright-time over-subscription ratios were 2.4 and 1.7, respectively. However, once one factors in weather and technical problems (e.g., the 16 day secondary failure in January 1996), effective oversubscription rates became 6.0:1 and 4.2:1, respectively, by the end of the semester.

Next consider the following breakdown of how we used the hours available to the WIYN queue experiment and then read the commentary that follows:

F95 Hours Available:    554 
Number of Hours Used
			Hrs     Percentage
Long Programs           222       40% 
2hrQ                     43        8%
Weather                 171       31% 
Technical Failures       92       17%
Misc Activity            26        5%

Notes:  Lost 8 (6 Dark/2 Bright) nights to secondary failure

                    Queue Completion Statistics
                      	 Number of Programs
                 	Input   Whole   Part    None 
Long Programs:
	Dark          	25      4       2      19 
Long Programs:
	Bright         	7      0       4       3 
2hrQ High Priority  	7      5       1       1 
2hrQ Best Effort    	9      0       1       8

Notes:

(1)     More dark hours than bright hours in a semester.

(2)     Lost more time to weather in bright time than dark time in F95.

(3)     WIYN Consortium using seven bright nights per month for T&E
	during F95, decreasing number of bright nights available for 
	science operations.

(4)     Programs took much longer to complete than expected, vastly
	exceeding allocation time. This was a combination of buggy
	telescope/instruments (this was the first real semester of 
	observing) and grossly underestimated overheads during the 
	time allocation process. Both problems have been vastly 
	mitigated for spring 1996.

(5)     Lost eight (6D/2B) nights to secondary failure

To put this in perspective, we received essentially identical numbers of 4-m and WIYN proposals, but NOAO only has 40% of the time on WIYN.

User frustration was heightened by two effects:

  1. We elected to give high priority to completing programs entirely before starting new programs during fall 1995. However, since observing overheads were unknown prior to proposal submission, most programs tended to require much longer to complete than anticipated during the TAC review. So, some highly ranked programs consumed more WIYN time than anticipated at the beginning of the semester (see "Active Long Program" tables at end of this message for details). Two changes in spring 1996 should improve this situation:
  2. We elected to "allocate" time to more programs than we could possibly complete to make sure we would have a good match between observing program and weather conditions at all RAs. The downside is most people expected to get some data although that was clearly impossible given the ground rules we were using. We tried to carefully match the number of nights available to the number of nights "allocated" queue time in spring 1996. While increasing the number of upfront disappointed applicants, it decreases the number of investigators who wait all semester and receive no data.

I have written about these issues in the March 1996 NOAO Newsletter (No. 45, p. 36) and I urge you to read that article. This article is also available on the Web (see the hotlink on the WIYN Observing Program Home Page (http://www.noao.edu/wiyn/obsprog/).

Users were also frustrated by lack of progress reports and clear information about how the queue was being executed. We have tried to reduce this frustration by creating the NOAO WIYN Observing Program Home Page (http://www.noao.edu/wiyn/obsprog/). This Web page contains links to application information, spring 1996 status reports (updated after every NOAO queue night), details about how the queue works, and queue experiment related NOAO Newsletter articles.

We know this experiment did not operate as smoothly or as satisfactorily as desired during fall 1995. Most of our problems were a combination of balky telescope (this was, after all, the first semester of science operations of a new, complex telescope) and inexperience on our part. We have fixed many of the telescope problems and have gained much experience. With the additional help of your feedback, we think that spring 1996 is going much better and we will continue improving things in the future. Any input you have on improving this experiment would be appreciated and considered carefully.

Nevertheless, no matter how efficient we become, it is a fact that WIYN is heavily over-subscribed right now. To get telescope time, whether through the queue experiment or as a classically scheduled observing run, you will need a strong science proposal.

Our highest goal remains to produce the best quality data possible. However, we are also committed to continuing to improve the level and quality of communication between us and you. We appreciate your concern and patience and we hope to serve you better in the future.


	     F95 Active Long Programs
			Hours
ID        Allocated     Used     Status 
2161        22.3        74.1     completed 
2164        34.0        52.3     completed/synoptic 
1929        42.5        25.3     partial 
2173        17.0        19.9     partial 
1999        25.5        13.7     partial 
1954         4.3        13.3     completed 
2006        25.5        11.9     completed/synoptic 
2085         8.5         3.5     partial 
2142        17.0         2.7     partial 
1937         4.5         2.0     partial 
Misc Cal     0.0         3.2

Totals     201.1       221.9

Notes:  One Typical 4-m Run = 25.5 hours
	Number of Equivalent 4-m Runs = 8.7
	


Back to the WIYN Observatory Home Page

Back to the WIYN Observing Program Home Page


Last Updated: 7/9/96, dsilva@noao.edu