[PREV - MODEST_PROPOSAL]    [TOP]

SOFT_ART

                                                         June 28, 2009     
                                                                           
  If you start with the notion that software
  is more of an art than it is a science,
  where does that get you?

      I've heard it suggested that this puts
      it beyond the realm of experimental
      science (and therefore there is no way
      you can scientifically evaluate a
      software methodology, etc).


      I don't quite follow why
      that should be so:
      certainly you couldn't
      conduct experiments to
      determine how to write         One complication, of course, is that
      Great Novels, because          "greatness" is somewhat subjective; but
      greatness is too difficult     I think the bigger problem is that the
      to measure.                    "test of time" critereon puts it beyond
                                     the scope of practical experiment.
      On the other hand, it would
      not be that difficult to get
      measurements of popularity,
      ala hollywood focus
      groups, or web click data.


      Comparing methods of writing Great Novels
      might be beyond the reach of science, but
      evaluating formulas for cranking out popular
      short-stories seems doable -- if perhaps not
      of any particular interest.



--------
[NEXT - COMPUTER_SCIENCE_PROOF]