[Haskell-cafe] Mozart versus Beethoven (was: Writing "Haskell For Dummies ...)

[Thorkil Naur]

Hello,

In the spring of 1978, I wrote a (circa) 700-word microprogram for 
multiprecision integer arithmetic on paper, typed it into a computer, had it 
cleaned of syntax errors by the micro-code assembler, printed it, and spent 
much of the summer in my mother's summer house debugging this program text by 
hand, without the use of any automated computing device of any kind. I found 
lots of errors, corrected them, rechecked the result by hand, found 
additional errors, corrected those and, finally, (in the autumn of 1978) ran 
the program for the first time. Every multiprecision integer operation but 
division worked. After some debugging, a single (rather silly) error was 
found in the division routine. I never found additional errors in this code.

This is not intended to imply that I am a Mozart rather than a Beethoven (most 
likely neither!) in the field of programming. Rather, it is an attempt to 
point out that the development environments that we use these days encourage 
a completely different mode of work than what was used some 20-30 years ago. 
Thus, today, I do like I have the impression most programmers do, compile and 
run (tests) as often as possible, even every very few keystrokes of code 
changes.

I am not an expert in the difference between composers like Mozart and 
Beethoven, but my expert father tells me that Mozart, reputedly, had a 
phenomenal musical memory that allowed him both to recall large sequences of 
music played to him and, undoubtably also, work with long sequences of 
"hypothetical" music, that is, music being composed, for prolonged periods, 
in his head, without the need to make any notes on paper etc.

It seems that such differences in modes of work does not imply any similar 
interesting or usefully utilizable difference in the way we should produce 
our programs. The analogy seems irrelevant, in other words.

Best regards
Thorkil

On Tuesday 12 December 2006 12:07, Kirsten Chevalier wrote:
...
> I've been thinking about this. Are there really any programmers who
> are like Mozart in the way you describe? Donald Knuth might be one, or
> at least, he wrote that he wrote and debugged all of TeX on paper
> before entering it into a computer and "only found 13 more bugs" (or
> something like that), once he did. I don't remember if it was 13
> exactly, but "13 more bugs" might be the closest that any programmer
> gets to Mozart, or at least any programmer in the 20th or early 21st
> century.
> 
...
> Cheers,
> Kirsten
> 
> -- 
> Kirsten Chevalier* chevalier at alum.wellesley.edu *Often in error, never in 
doubt
> "What is research but a blind date with knowledge?" -- Will Henry
> _______________________________________________
> Haskell-Cafe mailing list
> Haskell-Cafe at haskell.org
> http://www.haskell.org/mailman/listinfo/haskell-cafe
>