I've not used Python myself, so I tried doing some searches for benchmarks:
It showed interpretted Python being as much as 60 times slower than C++, and even the Psyco-compiled version was several times slower. Even places like python.org suggest using other languages such as C++ or Java for the performance-critical parts of the application.
I can certainly see Python's value as a scripting language, but writing a mud entirely in Python would seem to result in considerably worse performance than C++. Speaking as someone who had to spend a lot of time rewriting and optimising C++ code, after initially approaching my mud project with a "premature optimisation is evil" mindset, I can assure you that lots of apparently insignicant performance issues can really start to add up over the years.
I can't help but disagree - choosing the language is precisely the time when you should take into account potential performance issues (i.e., pick the right tool for the job). It's really not the sort of thing you want to discover two years down the line, when your code has become so slow that it requires a dedicated server you're not willing to invest in.
|