Human Factors

[utzoo!decvax!watmath!rhbartels]

The issue which Bernie keeps trying to raise has to do with
the double role UNIX is starting to play
   a) Hacker's delight and programmer's workbench
   b) Marketed system for naive users.
Flaming from standpoint (a) about real concerns in area (b)
is generating cross-talk static.
There are valid gripes to be raised about deficiencies in
UNIX for (b), but most of the readers of this newsgroup are
thinking in terms of (a), and are missing the point or are
not sympathetic to the point of (b).
Let the (a)'s thrash out their own issues and quit mucking up
the discussion of the (b)'s
Let the (b)'s clearly state their case and embark on a course 
of action.  In that regard, fellow (b)'s, what about putting
Ned Horvath on the spot?  
Dear Ned,
  What are the means of experimentation, measurement
  and evaluation in the Human Factors biz?
  What are the tools?  How do people go about setting
  up tests?  What constitutes a valid experiment?
  What statistics are taken?  How are conclusions drawn?
  How complex an environment can be usefully studied?
  How easy is any of this for the novice to try out?
  Does it really have any relevance to UNIX?
  Put up, or shut up.
                         Sincerely,
                         Richard Bartels