Priorities

Table 12: Do you prefer to use private programs to maintain a competitive edge?

#	Competitive edge	strongly agree			strongly disagree		n.r.
		1	2	3	4	5		mean
84		4	10	19	5	11	15	3.2 $\pm$ 1.2

On average, people were neutral in response to the question ``Do you prefer to use private programs to maintain a competitive edge?," see Table 12. One person wrote in that he had to (as opposed to preferred to) use private programs in order to do certain experiments at all.

Relatively few people filled out the section on priorities (see Table 13), probably because the layout was confusing and there were too many questions. The general response was roughly medium to medium-high priorities for most things. While the number of KEK users is low, they clearly placed a high priority on improving the KEK facility fitting program(s). Few people placed priorities on ALCR muSR software, probably reflecting the low usage of this technique. Is the ACLR technique being under-utilized?

Table 13: Prioritization of analysis software.

#	Prioritize for facility improvement	Lab	high			low		n.a., n.r.
			1	2	3	4	5		mean
85	fitting	ISIS	9	7	4	4	0	40	2.1 $\pm$ 1.1
86	fitting	KEK	6	2	1	0	0	55	1.4 $\pm$ 0.7
87	fitting	PSI	7	5	10	2	1	39	2.4 $\pm$ 1.1
88	fitting	TRIUMF	7	9	6	1	3	38	2.4 $\pm$ 1.3
89	Fast Fourier transform	ISIS	6	6	6	1	0	45	2.1 $\pm$ 0.9
90	Fast Fourier transform	KEK	1	2	4	0	0	57	2.4 $\pm$ 0.8
91	Fast Fourier transform	PSI	7	3	5	4	2	43	2.6 $\pm$ 1.4
92	Fast Fourier transform	TRIUMF	7	7	5	2	3	40	2.5 $\pm$ 1.4
93	integral- ALCR	ISIS	3	1	3	3	2	52	3.0 $\pm$ 1.5
94	integral- ALCR	KEK	3	1	2	0	0	58	1.8 $\pm$ 1.0
95	integral- ALCR	PSI	1	3	1	2	0	57	2.6 $\pm$ 1.1
96	integral- ALCR	TRIUMF	1	3	5	1	2	52	3.0 $\pm$ 1.2

People wrote in documentation, handbook and graphics as priority issues.

The last three questions (Table 14) set out to prioritize the general approaches of a universal data format, universal facility programs and private programs. The first two had mean responses of medium-high, with 68% and 69% respectively answering either high or medium-high, indicating strong support. Prioritizing private programs only got a neutral response.

The similar desire for a universal data format and a universal facility programs seems to come out of a general desire for making researchers' lives easier and more efficient, without necessarily a consensus on the best order to proceed.

In comments, some people saw no great advantage for a universal format as long as adequate conversion programs are available. Others complained that a given (ISIS, PSI) lab's current data format is not directly portable to operating systems other than VMS, without a data conversion. Some people asked for the ability to convert data to ASCII, to presumably ease programming demands on them. Application program interfaces (API) in C and Fortran for getting information out of data files might help people who need to make their own programs, as they won't need to know the exact details of the data format. Future data formats will have to be useable on all common computer platforms.

In comments, some mentioned the varying capabilities of different programs and that some useful features might be lost if a universal facility program were introduced. Also, programming your own is for some a learning experience and for others a necessity due to the demands of their experiment. Therefore, access to raw data directly will always be needed.

Table 14: Prioritization of universal data format and facility programs.

#	Prioritize for improvement	high			low		n.a., n.r.
		1	2	3	4	5		mean
97	universal data format	24	10	9	5	2	14	2.0 $\pm$ 1.2
98	universal facility programs	23	13	9	7	0	12	2.0 $\pm$ 1.1
99	private analysis programs	8	16	9	5	8	18	2.8 $\pm$ 1.4