🚀 go-pugleaf

In article <1992Nov5.205832.8919@linus.mitre.org> bs@gauss.mitre.org (Robert D. Silverman) writes:
>I think all of Herman's detractors are missing the point.

Don't lump me in with all of Herman's detractors.


>For Herman, the program IS NOT the product.
>The result of the computation is.
>Herman is not a programmer. He is a mathematician. Stop trying to pretend ...

Do you think I disagree with these statements?


>He probably spent a LOT of time on it...

He probably took a machine language routine and translated it at the
keyboard into something resembling C.  He never ran it through a
compiler, which could have saved us all much time and traffic.


>:the program is the first thing we look at, often the _only_ thing,
>Ah. Wonderful. So you never bother checking on whether the algorithm
>itself is well suited to your architecture.

Logic is a wonderful thing.  Sarcasm is not.


>:and provides the bulk of our understanding.  For us, a program is
>This is fine, but not for numerically intensive, and mathematically
>deep algorithms. How does checking the code tell you if it is the most
>efficient method for yor machine?

Herman admitted there were typos that would compile while breaking
the algorithm.  How am I supposed to tell if it _works_?


>I would call what everyone is criticizing Herman for:
>An overemphasis on form and structure, rather than content.

You completely miss my point, probably because you insist on seeing
me as part of a faceless horde.

The code Herman posted isn't efficient, it's BROKEN.  Will fixing the typos
hurt the efficiency?  How about adding comments?  How about whitespace?

And what is it doing on USENET, anyway?  So we can all ignore it?  No,
so we can READ it.  In this context, there is NOTHING more important
than making it READABLE.

Once again:  we asked for a program because it's the most concrete,
factual form of communication.  Herman gave us handwaving code.


Adam