Ben Okopnik <[EMAIL PROTECTED]> writes:


> As a small comparison, one of my students had a shell script that
> processed a pair of files, about 75MB each, by doing _something_ (I
> don't remember the exact details) that involved taking a line from file1
> and doing something to every line in file2 based on it, then moving down
> a line in file1 and repeating the process (note that this was done on
> some very fast and powerful Sun hardware.) According to his literally
> shell-shocked account on the third morning of a week-long Perl course,
> the time to run the script went from over 20 minutes to under 45 seconds
> when it was rewritten in Perl. Rewriting, by the way, only took him
> about ten minutes - the script was not very complex.

I, too, have engineered, written, debugged, and completed entire Perl
programs that completed before their original shell script
implementations did. There's something to be said for the right tool
and the right algorithm.

Michael R. Wolf
    All mammals learn by playing!

Reply via email to