[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Exposing internal functions

From: Jose
Subject: Re: Exposing internal functions
Date: Tue, 1 Oct 2013 09:42:32 +0300
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.0

Hello Kai.

Hello Jose,

scripts/optimization/private/__fdjac__.m seems to be a finite difference
jacobi operator used by the exposed functions
scripts/optimization/fsolve.m and scripts/optimization/fminunc.m. As I
don't know how you use that function, I cannot say if it makes sense to
modify your program using the exposed functions.

I have modified slightly fminunc.m, which, as you mention, uses __fdjac__.m. The new version (fminunc2.m) lives in a directory in octave's path, and also uses __fdjac__.m. I want to keep fminunc2 as similar as possible to the original fminunc, and therefore __fdjac__.m should be also in octave's path. So nothing esoteric.

__fdjac__.m only counts about 20 lines of m-code and there is no
matching function in MATLAB AFAIK. So I think that m-file it is to
trivial to be exposed. Maybe you can integrate the appropriate of the
two for-loops (each 9 lines of code) into your program logic directly
without cluttering your program?

I prefer not to do that, cause, as I explained above, I want to keep both versions as similar as possible.

At the moment I am using the symbolic link, which works fine. I guess I will have to live with this, if there is nothing better.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]