Python contains many very-high level data types, such as lists and dictionaries. However, it does not contain an array type that behaves in the same way as matrices in mathematics, i.e. operations take place element-wise2.2.
The Numeric Python library adds this feature. It is implemented in C for speed and is highly optimised. It can be used as a drop-in replacement for IDL and Matlab in scientific research[#!numpy!#].
The version of Numeric Python used during the trial was 21.0.