Saturday, December 12, 2009

My dear Matlab

I feel like I am doing research in the past two weeks. Basically trying to write a program to do data fit based on a certain hypothesis.

Doing numerical analysis reminds me of the days of Matlab. Matlab is definitely the way to go in these kinds of things. I’m doing it on c# which is not as convenient as Matlab. You can plot things on the fly with Matlab which is nice for both debugging and visualization. Luckily, my company has this 3rd party charting software which makes it possible to plot my data but it’s not on the fly like Matlab does. It can probably do it on a window app basis and can hence be done on the fly but we have been using it strictly for web purposes and I don’t have the luxury to invest any time in investigating how to use it in window app.

But without a doubt, Matlab is still way more superior in numerical analysis with a lot of great built in functions and easy matrix manipulations. I never really appreciated Matlab until now. I always wonder y all the researchers are using Matlab while there’re so many other better supported software out there. Well… now I know y. Basically, it took me a few days to do something that 詹占 could do in Matlab in less than half an hour.

And again, I’m glad that we have some sort of charting ability. Otherwise, I either have to spend time finding some free app that I can integrate into my thing or export all the data in excel and chart in excel. So I guess it could be worse.

Oh… I miss Matlab. (totally forget how to write code in Matlab though...)

1 comment:

Michael Croucher said...

MATLAB is very cool but if you can't get it then there are all kinds of non-MATLAB solutions that will do the kind of thing you are talking about. For example:

Python with matplotlib
http://matplotlib.sourceforge.net/

Octave (largely compatible with MATLAB)
http://www.gnu.org/software/octave/

Scilab
http://www.scilab.org/

Cheers,
Mike