Numpy does have much better scientific capabilities, so you can use vanilla Python for the basic coding and Numpy routines when you need them. APL uses arrays

Numpy does have much better scientific capabilities, so you can use vanilla Python for the basic coding and Numpy routines when you need them. APL uses arrays for everything though. I just wish it had taken off more on the scientific front instead of the financial side as the linear algebra stuff isn't nearly as fleshed out as Matlab/Numpy/Mathematica.