[PREV - CASE_STUDY_2]    [TOP]

CASE_STUDY_3


                                             June 22-27, 2007


  The Plat_Forms contest!

  Someone is actually
  trying to get real data!

          [ref]
          [ref]
          [ref]
          [ref]


  Unlike my scenario (using
  undergraduate volunteers
  in a university setting),
  they're working with
  small teams of volunteers         This selects for people who
  in a marathon-contest setting.    think they're good.  These are
                                    not "average" programmers.
                                    The teams volunteer as teams,
                                    so presumably inter-personal
                                    dynamics have been worked out
                                    already before the coding
                                    starts.


I love these guys.  They sound like
they've got some funny biases --
they're corporate-weenie java-lovers     Problems:
at heart -- but they're actually
working out ways to get *something*      It's not such
like scientific data comparing           a mystery that
different "languages" (really,           WDSL support in
"worlds"/"cultures") of software         perl is weak,
development.                             the mystery is
                                         why anyone
  The actual notable thing               bothers with it.
  about these results (to me)
  is that they're all so                 They've chosen an
  similar.                               easy, high-level
                                         way of detecting    If you knew in
  Three decent programmers               possible SQL        advance how the
  who are familar with their             injection           test was going
  tools can put up a social-             vulnerabilities,    to be graded,
  network site in a few                  but it yields       you'd fix the
  weeks of work.                         false-readings.     software to get
                                                             it to act that
    These results might help                                 way.
    you choose one or the
    other language, but they
    get awfully close to proving
    that "languages don't matter".



      A limitation:

      By design, they're working
      on a problem in a well-
      developed, mature field.

      If you were trying to do
      something new, would there
      be clearer advantages?

          Also, the "30 hour marathon"
          is a little artificial. But
          some software start-up jobs     Does it provide an
          are like that, at least.        edge for youth over
                                          wisdom?


--------
[NEXT - CASE_STUDY_4]