I have the following comments on SAS Institute's Linux Interest Survey (http://www.sas.com/misc/linux/). As there is no response contact provided, I'm bouncing comments to the SAS-L and SAS-Linux discussion lists.

1). Provide a contact at SI for more information and/or survey feedback. There are both errors and aspects of the survey which could stand improvement. A 'comments' field would be helpful. I've received a number of insightful (and occasionally inciteful) tidbits through the free-form fields on my own survey.

2). Among the organizational size and purchase authority questions, it might make sense to ask what sort of organization the respondent represents. I suspect that Linux interest will be highly bi-polar, with consultants such as myself looking for a way to maintain SAS/Unix / SAS/Linux skills on personal equipment, and companies looking to deploy SAS on Linux in a production environment. These are two wildly different markets, and should be approached differently.

3). Hardware interest question is ambiguous and incomplete. Does 'Intel' represent x86 architecture only, or include Merced? Linux also runs on Mips, Mac, and a growing list of other chips. Embedded systems might be yet another area of interest.

4). Distributions. It's "Debian", not "Denebian" (unless I'm failing to keep up with my distros). I'd also add SuSE and TurboLinux to the list (German and Asian distributions, respectively). Given Linux politics, I'd strongly recommend ordering the list alphabetically to avoid charges of favoritism.

5). "have you made kernel modifications" is unclear. Does this mean "have you modified kernel source code?" (no, I haven't) or "have you compiled a custom kernel for your system?" (yes, of course).

6). "would you be willing to use a specific kernel?" is unclear. Does this mean "would you be willing to use a specified release or version number of the Linux kernel?" (probably) or "would you be willing to use a specific compiled kernel binary?" (possibly not, though it would be fairly easy and cheap to set up a dedicated SAS/Linux server, in which case, yes, I'd consider it).

7). Desktop environments -- I'm surprised that Motif and CDE don't make the list, if only to show how unpopular they are among Linux enthusiasts. I'd also expect to see a 'fvwm/fvwm2' option -- this is an mwm-like window manager which is pretty standard fare among Linux fans. Being an advocate, I'd recommend a WindowMaker option (great, lightweight WM, GNOME and KDE compliant, http://www.windowmaker.org/). A general solution to listing each of the many Linux window managers would be to simply ask whether the respondent (or her organization) uses an integrated user environment (IUE) or just a window-manager.

7a). The technical bonus question would be to ask what GUI toolkit libs are installed on a system, should be a 'check all that apply' option. Information may be erratic, but it may help tune SI's attempts to identify a good target for future GUI development. Motif loses badly, even in its Lesstif implementation. Much more appealing, stable, and popular alternatives are available.

8). Personal information. If SAS wants to contact me, well, they know where to find me. However, it would make sense for the survey to collect, independently, respondent contact info, _and_ contact information for other key roleplayers in an organization. It's not clear whether or not "please specify" means I should provide my own contact info (not previously recorded) or only that of another person. As it is, the survey captures the respondent's name and title, but not a phone, email, or postal address. This is silly.

8a). There doesn't appear to be any attempt at response validation. This concerns me from two perspectives: it makes survey results interpretation much more difficult, and it makes the task of defending these results within SI's organization more complicated. I'd feel more assured of SI's committment to exploring Linux as a viable platform if the instrument requested more of me, rather than less.

9). SAS is more than welcome to crib what they want from the interest inventory I compiled (http://pweb.netcom.com/~kmself/SAS/sas_user_survey.html). As SI have taken the effort to compile their own survey, I encourage people to respond to the SI survey. I am interested in sharing my results with SI, if the appropriate SI staff will contact me (again, there appears to be no contact information provided for SI Linux efforts. I assume that Open Systems staff are involved).

-- Karsten M. Self (kmself@ix.netcom.com)

What part of "Gestalt" don't you understand? Welchen Teil von "Gestalt" verstehen Sie nicht?

Web: http://www.netcom.com/~kmself/ SAS for Linux: http://www.netcom.com/~kmself/SAS/SAS4Linux.html