->The benchmark test suite consists of programs written in: ->C, Java, Python, and a hybrid (consisting of a main routine ->in Python with a C subroutine). Here are the results... -> I would be interested to know what versions of the various interpretors you used are as well as the OS & load you ran the tests under.
An inital glance the benchmarks look about right for how the various languages operate. While I don't agree with you on the conclusion that Python should be used as a prototyping language you need to take into account the amount of effort needed to do something 'useful' in a particular language. C is almost never the best choice these days (and I'm a huge C fan) as to get anything complex done it takes lots of effort. C++ is a bit better as it provides for some primitive 'Generics' and Java is a 'jack of all trades' lanugage. The various scripting languages provide few neat features that don't exist in the above languages such as the code=data paradigm of the 'eval' function in perl, garbage collection (yes Java has it but its pretty much a rogue feature you can't control) and weak typing. Their cost is obviously brute speed though machines these days are 'fast enough' to ignore preformance on most (not all) problems.
Enough of my rant, congratulations on a great post I really enjoyed it.
That is all,
-- Sean "I'm pretty sure this is true . . . I wouldn't bet my life on it, but I'd bet yours." - Erik Guentner