Can anyone point to serious comparison of Python runtime footprint versus Java?
I can't compare memory footprint because it really depends on classes what you load/use. But what I can tell you that Python (IronPython 2.7 in particular) has real memory leak problems. Especially with third party well used ones like Financial.
When Java application/server runs without issues with rare cases which could be identified with common tools Python grows in memory constantly.
Memory dumps shows that Python itself as well as most of packages don't pay attention for common classes like String and keep them in different parts of the execution modules. It is hard and unwise to go through all these sources and fix all leaks.
I was trying a lot to fix the issues but finally gave in and simply restart application when it reaches some memory threshold.