Re: Development: A Structured Problem Area? (Background on Dreyfus)

On Tue, 2004-06-15 at 22:40, Derek Richardson wrote:

> However, if
> pattern-recognition is merely the way good programmers function, but 
> rule-based
> systems can adequately approximate the results of good programmers, then 

Ah, but there's the ruby.  According to other studies (that I cite in
the Dreyfus talk), while novices require context-free rules in order to
operate and be succesfull, sticking to rules-based behavior *limits you
to a novice level of effectiveness*.

Depending on who you believe, good programmers are 10x-20x more
productive and "better" than poor programmers.  If a CMM-styled
automation program (as you put it) can only deliver us a bunch of
1X'ers, then what's the point?  We want the pattern-recognizing, good
programmers, the 20X-ers. 

Do you know how Dave and I got introduced to Dreyfus?  My wife is a
registered nurse, with a Master's degree in Nursing Informatics.  She
told me about the problems the nursing profession had in the late 1970's
and early 1980's.  As I started digging and researching the stuff, it
started to sound really familiar.  If you think about it, the medical
profession in general is faced with some very sticky problems.  You need
a basic (and large) body of knowledge with which to operate, but the
rules only take you so far.  Every patient is different, reacts
differently, and there are millions of hard-to-quantify variables that
nearly guarantee chaotic behavior.

The skilled doctor -- or nurse -- can look at a patient and make an
advanced diagnosis that only tests can confirm.  And they can't
necessarily explain how they knew, it's just that "the patient didn't
look right".  That's what you label as the "pattern-recognition" model,
based on their years of expierence and intuition.

How do they "automate" their process?  They don't.  You become a doctor
(or a nurse) by practicing.  You go to school forever, and then you get
to watch and maybe assist a doctors for a long time.  You are an
apprentice while you gain the necessary real-world, hands-on
experience.  That's the only way you can learn this complex field.

But of course, we don't do those sorts of things.  We pretend that
sheep-dip style education is sufficient and that you can practice for
the first time on the live patient (usually a mission-crticial one, at
that).

But back to your question, I think the whole damn thing is non-linear.

/\ndy



------------------------ Yahoo! Groups Sponsor --------------------~--> 
Yahoo! Domains - Claim yours for only $14.70
http://us.click.yahoo.com/Z1wmxD/DREIAA/yQLSAA/nhFolB/TM
--------------------------------------------------------------------~-> 

 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
     http://groups.yahoo.com/group/pragprog/

<*> To unsubscribe from this group, send an email to:
     pragprog-unsubscribe-hHKSG33TihhbjbujkaE4pw@xxxxxxxxxxxxxxxx

<*> Your use of Yahoo! Groups is subject to:
     http://docs.yahoo.com/info/terms/