On 04-Oct-2004, Jon Cast wrote:

Fergus Henderson held forth:

...

If analyzing the performance and space usage of your programs is important, then Haskell may not be the best choice of language.

I disagree. If performance and space usage are sufficiently important, Haskell may not be best. But it's really not that much easier to analyze performance or space usage under strict languages.

You're welcome to your opinion, but I don't agree; I find it much easier to analyze performance and space usage for strict languages.

In either case, the golden rule is: profile. Reading the source code will tell you very little.

Profiling can only tell you about the costs for a finite set of test cases. If you want to write robust and reliable programs, you need to analyze the worst-case memory usage, and that requires looking at the code, not just profiling. Of course, you can try to find a representative test set which will exercise the worst-case behaviour, but that task of determining which tests will exercise the worst-case behaviour will itself requires analysis of the code. That analysis is IMHO much more difficult in lazy languages. -- Fergus J. Henderson | "I have always known that the pursuit Galois Connections, Inc. | of excellence is a lethal habit" Phone: +1 503 626 6616 | -- the last words of T. S. Garp.