I'm aware of his contributions to computer science but he did not program -
writing down algorithms with pen and paper is fine and useful but it avoids
the nitty-gritty problems of actual programming.

>From Wikipedia: Dijkstra was also noted for owning only one computer (late
in life) and rarely actually using them, in keeping with his conviction that
computer science was more abstract than mere programming...

He expresses the disdain for the practical difficulties of actually getting
a program to work correctly that is all too common in computer science.  At
the recent APL conference, a (non-CS) college professor told me that a
computer science student in his class argued that he should get a passing
grade for his computer program because it ran to completion - never mind
that his answer was off by orders of magnitude.  That same day, another
(non-CS) professor also mentioned to me how the CS students were usually bad
programmers.

There's an old joke that goes like this:  in three adjacent rooms of a hotel
are sleeping an engineer, a physicist, and a computer scientist.  In each of
the rooms, a fire breaks out in the waste basket near the bed.

The engineer wakes up, sees the fire, and pours a pitcher of water on it to
put it out.  Just to be on the safe side, he refills the pitcher and douses
the ashes.

The physicist wakes up, sees the fire, calculates from his estimate of the
heat being produced just exactly how much water he needs to pour on the
flames, pours this amount and douses the flames.

The computer scientist wakes up, sees the fire, sees the pitcher of water by
his bed, decides it is a solvable problem, and goes back to sleep.

None of this detracts from the fact that Dijkstra was right about GOTO being
harmful.  One of the things I originally liked about APL was how it avoided
much unnecessary looping - probably the branch arrow was chosen because it
was the simplest, most general construct and it was often not necessary.

On 11/6/07, Boyko Bantchev <[EMAIL PROTECTED]> wrote:
>
> On Nov 6, 2007 6:03 PM, Devon McCormick <[EMAIL PROTECTED]> wrote:
> > I find it interesting that he talked a lot about the practice of
> > programming but did not do it; this might help explain his blind spot
> > toward APL.
>
> He did practise a lot.
>
> In his younger years, Dijkstra was one of the (two) creators
> of the very first compiler of Algol 60 (after being one of the
> designers of that language), and led the design and creation
> of the very influential ``T.H.E.'' operating system.
>
> Although his later work is not directly related with
> large-scale programming, it contains lots of small but far
> from trivial algorithms which he wrote mostly in a pseudocode
> notation of his own.
>



-- 
Devon McCormick, CFA
^me^ at acm.
org is my
preferred e-mail
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to