Re: The fabled "Matlab compatibility"

[Daniel Carrera]

Hi Jordi

On 8 November 2012 23:26, Jordi Gutiérrez Hermoso <address@hidden> wrote:

These are a few more:

    http://wiki.octave.org/FAQ#What_features_are_unique_to_Octave.3F
    http://www.gnu.org/software/octave/doc/interpreter/Broadcasting.html

Nice. I really like the one about defining functions by entering code directly on the command line. As someone who uses Fortran 95, I also appreciate the endfor, endwhile, etc. rather than the dangling "end".

There are a few other differences in some functions here and there. We
do not shy too much from making Octave functions do more if we find it
useful and doesn't break existing Matlab code. For example, the norm()
function accepts a "columns" parameter to do something that is
frequently desired but Matlab can't do this way.

Sometimes we have to backtrack when we do this. For for instance, I
had to recently modify the perms function which used to take a 2nd
argument. Matlab then added a second argument which did something
different than what we did. Thankfully, nobody seemed to be using our
version of the two-arg call, so it wasn't a big problem to change it
to look like Matlab.

Would it be fair to say that syntax-wise, Octave is aimed to be a superset of Matlab? So ideally all Matlab code is valid Octave, but within that constraint Octave will add improvements wherever it sees an opportunity?

Cheers,

Daniel.

--
I'm not overweight, I'm undertall.