🚀 go-pugleaf

In article <1991May9.004140.1116@tc.fluke.COM> kurt@tc.fluke.COM (Kurt Guntheroth) writes:
>I can't believe all this <expletive> about the programming contest.  How is
>football or basketball like real life.  It's a CONTEST.  Y'know, a GAME?  An
>artifical (and in fact unnatural) set of constraints placed on the
>contestants for the purpose of making the contest lively and exciting?

The Obfuscated C Code Contest is a game, run for fun.  The ACM exists to
further the computer industry, and I think many of us assume that the ACM
programming contests are intended to be related to the education of future
computer professionals.  It's possible to design contests that are both fun
and relevant.  At MIT, there's a design class in the Mech.Eng. department
that has a contest each term, where students are given a collection of
random objects (strings, spools, a motor, etc.) and must build a device
that will participate in a head-to-head competition against a classmate's
device; the real-world relevance of this is to the design of devices with
constraints on the raw materials.  Many schools have "egg-dropping"
contests, which are similar.  In both cases, the particular problem is
unnatural (no egg distributors seem to be replacing their cartons), but the
skills they exercise are precisely those which are used in the real world.

The programming contests, however, do not exercise many of the skills that
are considered important to learn.  In contrast, I think they test the
talents that are somewhat inborn, and which caused us to gravitate towards
the programming field, rather than that skills which must be learned in
order to apply these talents best in the real world.  There are many
extremely prolific programmers out there who aren't able to work
effectively as part of a team, whose code is virtually unreadable, or
who are very sloppy about error handling, etc.

There was a time when the hallmark of a great programmer was his ability to
produce lots of programs or to squeeze every last millisecond off the
program's runtime.  I'd like to hope that we've progressed beyond this, but
the programming contests still reward these same things.

Of course, one problem the contest designers have, and which is shared by
industry, is that objective criteria are necessary to reward the "best"
programmers.  While many of us have progressed beyond the above measures of
good programming, we've had a hard time coming up with objective ways to
measure the kinds of qualities that we do want out of programs and
programmers.

-- 
Barry Margolin, Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar