Have you figured out an easy way to publish your MATLAB projects for public consumption?

I've come up with several MATLAB scripts that I'd love to be able to make available for public use. As far as I can tell, MATLAB is not built for this. There's no way for me to just embed the code into a web page or package it into a standalone app. (yes, I'm aware of the Compiler, but it's not available for individuals)

It seems like the intention is research and development. Then, if you want to share it publicly with someone who doesn't have MATLAB, it has to be translated to C++ or Python or JS or something similar.

Am I understanding things correctly?

[Walter Roberson]

The compiler is available to individuals -- in the sense that the Standard license is available to whoever wants to pay for it (and is not in a country under sanctions.) The compiler is not available to Home or Student licenses, however.

[Nathan Lively]

Seems funny to limit it that way. Wouldn't Mathworks like me to be sharing the things I'm creating and promoting their product?

[Walter Roberson]

Historically, Mathworks has encounted problems where purely commercial enterprises has purchased the lower-cost licenses and deliberately violated or ignored the license conditions.

I believe there is concern that people would use Student or Home licenses to generate executables for commercial purposes.

I do not have any information about how highly that particular concern ranks against other concerns.

[Nathan Lively]

Interesting. Thanks Walter!

Answer 1

[Rik]

(this is not a drop-in solution)

You could bundle your m-file with a copy of GNU Octave.

The main upside is that it will be free (under many circumstance, you should check the licenses of both Matlab and Octave).

The main downsides are that it will take developer time to confirm everything works as expected (and create work-arounds for when it doesn't), the debugging tools are much worse, the performance is much worse, and the support is much more hit-and-miss.

I personally make sure my code will run on very old Matlab releases, as well as Octave. I would estimate implementing the workarounds takes about the same time, mainly due to the debugging tools. (very old Matlab in this case being Matlab R13 (v6.5), which was released June 2002)

For Linux it seems to be possible to have the OS treat your functions as a stand-alone application, but for Windows this doesn't seem to be possible (yet?). MacOS is probably even further away.

[Nathan Lively]

Thanks Rik. I forgot about Octave. That's a great tip. My goal is to get friends to test the things I'm building so I can iterate and improve faster, so that might be a way to do it.