[cgranade]

That's bound to throw off the Java and Python numbers quite a bit. JDK 1.2.1 was problematic, and I think a lot of its poor performance there could probably be fixed with no more than a simple recompile. The past 10 years have also seen a wide diversification of Python implementations, including IronPython, Jython and the JIT-based interpreters. It would be very interesting to compare the performance of the same programs across platform implementations to try to knock down the variations and get to the effects of the choice of language itself.

[kllvql]

Both of those languages have seen vast improvements in speed throughout the past ten years. Java 1.4.2 was immensely faster than any previous Java implementation. I would be very interested in seeing this study redone with the newest Java, Python, Ruby, and something Lispy.