[IPOL discuss] Python as a platform for reproducible research (... not!)

[Miguel Colom]

Quoting Ofer <ofer.bartal at weizmann.ac.il>:
> If c++ code also needs to be maintained, then do you think c++ is easier to
> maintain than python code? Lets focus on the language and assume for the
> discussion that the c++ libraries allowed by IPOL exist and are stable in
> python.

In my opinion, once the code is written, both C/C++ and Python sources  
would cost the same in terms of maintenance.

The main problems I see are:
1) Python is very slow.
2) The existance of long-term stable versions of the libraries for  
numeric computation available in Python. For C/C++ we could find some,  
but I'm not so confident with Python. The excerpt Nicolas showed is a  
good example.

If we could find long-term stable libraries for numeric computating  
and parallelization of the execution, and as well a way to efficiently  
clusterize many machines or using a NUMA architecture to bypass the  
slow execution of Python, I'd strongly recommed the use Python as the  
main language.

But for the moment, this hasn't arrived and C/C++ is still the best  
option for us.