SEUL: Developing a user interface

["Laura A. Tisoncik" <muskie@grrltalk.net>]

Now that I'm semi-back (I'll skip the details, but they're very ugly,
and Linux wasn't at fault: apologies for defaming *The* operating system
in my frustration)... 

Few if any of us are representative of the end users we target. In order
to develop a viable distribution for ordinary end-users,I don't think we
can rely on our impressions, speculation, and personal preferences. We
need organized, empirical data coming from real-live end-user testers.

One way to get that is to start out with a pre-existing distribution or
an initial SEUL base distribution, a few test machines, a collection of
willing (or at least coercible) testers, and a basic task list. Starting
cold, ask the testers to try to run through the task list. Run these
tests every few months as the distribution evolves, using different
testers to eliminate the learning factor. Get the feedback, analyze it,
and, where possible, incorporate the results into the design of the
distribution.

I may have a machine I can dedicate to this within the next month or
two, and, assuming the machine materializes, I'm willing to kidnap
unwitting friends and neighbors for experimentation. 

A) Any opinions on this?
B) Any suggestions for a basic task list to be tested?
C) Anyone else willing to participate in human experimentation on their
loved ones? :)

---
Laura A. Tisoncik
muskie@grrltalk.net