Re: Octave and Database connectivity

[fork]

Philip Nienhuis <prnienhuis <at> users.sf.net> writes:

> Maybe that's the natural inclination if you're a mathematician or software
> engineer. But I see no reason to be that negative about Matlab and its OSS
> siblings in other settings.

I would second this opinion.  I have programmed extensively in Matlab, R,
Python+numpy, and for math programming on any scale, Matlab/Octave is by far my
favorite language.  While it has deficiencies (I especially miss Python's object
model, named parameter syntax, and dict functionality), ML/Octave is far better
in many respects (I HATE Python's zero based indexing and indentation rules, and
I hate everything about SPLUS/R except for the extensive libraries).

I skimmed the blog post, and I must say I disagree about a few fundamentals --
Matlab/Octave has very flexible data structures (cell arrays + structs), plus I
think vectorization is actually a good thing (it helps you learn to think like a
mathematician and loops are verbosely ugly).  I sort of like Python's module
system, but I also really like the quick access to functions as they are being
edited in Matlab's system (tradeoffs, tradeoffs).

I think the only problem with database connectivity in Octave is the lack of a
good library.  Ditto for web app programming.  I would try writing scripts as
specified by Sergei and others and see if that is good enough before optimizing
prematurely.  Those scripts could be written to generate binary formatted data
for quicker loading, I would imagine (sed is your friend).

One language to check out is Julia.  I don't like it because vectorization is
ugly in it.

Random thoughts, best wishes.