On Sat, 2014-12-27 at 13:46 +0000, aldanor via Digitalmars-d-learn wrote: > On Saturday, 27 December 2014 at 10:54:01 UTC, Russel Winder via > Digitalmars-d-learn wrote: > > I know much less about R, but the whole Python/NumPy thing > > works but > > only because it is faster and easier than Python alone. NumPy > > performance is actually quite poor. I am finding I can write > > Python + > > Numba code that hugely outperforms that same algorithm using > > NumPy. > There will sure be some algorithms where numba/cython would do > better (especially if they cannot be easily vectorized), but > that's not the point. The thing about numpy is that it provides a > unified accepted interface (plus a reasonable set of reasonably > fast tools and algorithms) for arrays and buffers for a multitude > of scientific libraries (scipy, pytables, h5py, pandas, scikit-*, > just to name a few), which then makes it much easier to use them > together and write your own ones.
Agreed, it is not NumPy that is the win, it is PyTables, Pandas, SciKit-Learn etc. These are the standard tools because they are domain specific and aimed at the audience. The audience neither knows nor cares that NumPy is actually not very good because they have the tools they need and nothing to compare them against – unless Julia gets real traction, or a language like D can use it's one or two entries in the field to create a usable set of libraries. As with the Vibe.d, and Dub experience, pick a field, write and use something that does the job better than anything else in that field, then market the experience. -- Russel. ============================================================================= Dr Russel Winder t: +44 20 7585 2200 voip: sip:[email protected] 41 Buckmaster Road m: +44 7770 465 077 xmpp: [email protected] London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
signature.asc
Description: This is a digitally signed message part
