views:

151

answers:

2

Criteria for 'better': fast in math and simple (few fields, many records) db transactions, convenient to develop/read/extend, flexible, connectible.

The task is to use a common web development scripting language to process and calculate long time series and multidimensional surfaces (mostly selecting/inserting sets of floats and doing maths with them).

The choice is Ruby 1.9, Python 2, Python 3, PHP 5.3, Perl 5.12, or JavaScript (node.js).

All the data is to be stored in a relational database (due to its heavily multidimensional nature); all the communication with outer world is to be done by means of web services.

+10  A: 

I would suggest Python with it's great Scientifical/Mathematical libraries (SciPy, NumPy). Otherwise the languages are not differing so much, although I doubt that Ruby, PHP or JS can keep up with the speed of Python or Perl.

And what the comments below here say: go for Python2.6. This has mature versions of all needed libraries, and if you follow the coding guidelines, transferring some day to Python 3 shouldn't be a too big pain.

Peter Smit
I think numpy has no release for python 3, so I would recommend python 2
Fabian
same goes for the DB drivers -- few are ready and mature for version 3
Marco Mariani
+4  A: 

The best option is probably the language you're most familiar with. My second consideration would be if you need to use any special maths libraries and whether they're supported in each of the languages.

tooba