One disadvantage of distributing CMU Common Lisp by anonymous FTP, without
any licensing hassles, is that we have no way of knowing how many users we
have and who they are. We need this information for many reasons, as do
the funding agencies who keep us going. So I must ask all current users
of CMU CL (where "user" means something more than "tried it once") to take
a few moments, fill out this survey, and E-mail it back to me. My internet
address is ·····@cs.cmu.edu".
The best way to respond is probably just to copy the survey part of this
message into an editor buffer and insert your answers between the
questions. Free-form answers are OK -- these responses will be read by
humans. Be as brief as possible, but if you have a complex opinion on some
issue, take whatever space you need.
If you can speak for a whole group of people, that's great -- just send one
message, preferably from one designated person. Please only respond on
bahalf of the part of an organization that you really know about. Don't
try to respond on behalf of all of NASA, for example, unless you really
know about CMU CL usage in all of NASA. If you're not sure whether someone
else has counted you, please respond; we would rather get some overlapping
responses than have people fall through the cracks.
While we are taking this census, we may as well gather some information on
what your priorities are for improvements and extensions to CMU CL. This
will guide us in setting priorities within our project.
After this survey simmers down, we'll have a *separate* survey for people
who would like to use CMU CL, but who can't because it doesn't run on the
right machine or because it lacks some essential feature. This will also
guide us in setting priorities. But, please, do not respond now if you're
just a would-be user of our system.
We will probably repeat these surverys every 4-6 months. We appreciate
your help in this effort. This information should help us to serve you
better.
Scott Fahlman
CMU Common Lisp Project
School of Computer Science
Carnegie Mellon University
Pittsburgh, PA 15213
===========================================================================
CMU Common Lisp Users Survey, February 1992
1. Your name, organization, and location.
2. Number of CMU CL users represented by this response. (If this does not
cover the whole organization listed above, please indicate the group
you speak for.)
3. Is your group funded by DARPA, NSF, of some other agency of the United
States federal government? (If so, please specify.)
4. Very briefly, what are you using CMU CL for? (For example: medical
expert systems, development of AI tools, general hacking, teaching Intro
CS, secret or proprietary uses...)
5. We are thinking about eliminating support for the built-in Hemlock
editor and devoting those resources to improving the Gnu Emacs
interface. Good idea, bad idea, or don't care?
6. What user-interface toolkit, above the level of CLX, would you like to
use if it were bundled with CMU CL for free, with decent performance and
support? (CLUE, Garnet, CLIM, LispView, CLM/Gina, Picasso, other,
none...)
7. Would you be interested in support for some three-dimensional data
visualization package from CMU CL? If so, which one? (PHIGS/PEX,
AVS, Explorer, other...)
8. We will definitely be producing a generation-scavenging GC, a
tree-shaking purifier for delivery of small systems with only the
runtime support needed by a given application, a stepper and improved
debugging interface, and a better call-out to C. We have longer-term
plans to produce a much more efficient and complete CLOS implementation.
What else do you want? (Please indicate the priority you would assign
to each suggestion.)
9. We currently run on Sparc/SunOS, Sparc/Mach, IBM RT/Mach, and
Decstation/Mach. We expect CMU CL to run on Decstation/Ultrix for free
as soon as DEC releases the OSF-compliant version of Ultrix. What other
machine/OS combinations would be of greatest use to you or people at
your site? Would this port bring in new users or enable new
applications? (Note: A port to DOS or the native Macintosh system is
probably out of the question, though we might be able to port to Unix on
either machine. RISC machine ports are much easier.)
10. We are thinking of porting CMU CL to run on a supercomputer or
large-scale parallel processor. Do you have any computation-intensive
or number-crunching applications that would benefit from being done in
Common Lisp (with suitable extensions)? If so, please describe.
11. How would you rate the relative importance of the following: Ports to
new machines, better CLOS performance, better general performance, new
features?
12. Other comments, compliments, complaints?