[Numpy-discussion] multi-dimensional c++ proposal

Sturla Molden sturla.molden at gmail.com
Tue Oct 28 08:02:57 EDT 2014

Neal Becker <ndbecker2 at gmail.com> wrote:

> That's harsh!  Do you have any specific features you dislike?  Are you objecting 
> to the syntax?

I have programmed C++ for almost 15 years. But I cannot look at the
proposed code an get a mental image of what it does. It is not a specific
feature, but how the code looks in general. This is e.g. not a problem with
Eigen or Blitz, if you know C++ it is not particularly hard to read. Not as
nice as Fortran or Cython, but ut is still not too bad. Boost multiarray
suffers from not being particularly readable, however, but this proposal is
even worse. I expect that scientists and engineers will not use an
unreadable array API. When we write or maintain numerical algorithms we
need to get a mental image of the code, because we actually spend most of
the time looking at or reading the code.

I agree that C++ needs multidimensional arrays in the STL, but this
proposal will do more harm than good. In particular it will prevent
adoption of a usable array API. And as consequence, it will fuel the
problem it is trying to solve: C++ programmers will still used homebrewed
multiarray classes, because there is no obvious replacement in the standard


More information about the NumPy-Discussion mailing list