Bruce McFarling wrote:
> On Mar 25, 4:11 pm, JK <jk.use...@[EMAIL PROTECTED]
> wrote:
>> Paul E. Bennett wrote:
>
>>> Only those who think complex problems demand complex solutions think
that
>>> way. By all accounts this is a large number.
>
>> I find statements like this curious. If you have a "complex" problem,
>> and there exists a simple solution, then in what sense is the problem
>> complex?
>
> It may make more sense when elaborating "demand complex solutions"
> into "demand complex problem solving approaches".
Oh yeah. A perfect example: I was for a while on a committee
that developed the standard for communicating with certain kinds
of roadside devices. When I joined the committee, it had just
finished version 1.0 of the standard, which was about 30 pages
long and quite easy to get a grip on (both physically and
mentally). A new chairperson had just taken over, however, and
for some reason this individual insisted that (1) the "old
methodology" of a bunch of industry experts who really knew the
domain, sitting around a table and banging out the standard
language in a somewhat ad-hoc manner, was not adequate -- we
needed a real "systems engineering approach", with "user needs"
and "traceable requirements" and so on. And (2) the form of the
standard document itself had to reflect this process, by including
(in three different formats) explicit explanations of the mapping
between user needs, requirements, and normative language in the
standard. The result is that version 2 of the document, which took
two years to produce and increased the supported functionality by
about 20%, was over 250 pages long! I don't think anyone bothers
to read the thing anymore. (But, y'know, when Federal funds are
available, we gotta use 'em.) I left the committee when I realized
I was actually ashamed of the product it produced, and I didn't
feel it was in my power to change the situation.
> The complexity of the simplest possible solution is, I would say, a
> good measure of the complexity of the problem. The idea that the more
> complex problem demands that a more intricate *approach* to problem
> solving means that the complexity of the solution risks being
> amplified by the complexity of the problem solving technique.
>
> "As simple as possible" may be fairly involved for a particularly
> complex problem, but as a rule of thumb, its better than "As complex
> as you can cope with".
Agreement.
--
(declare (antichrist i) (anarchist i)) ; -- the sexp-pistols


|