I was a hardcore matlab programmer for 10 years. I discovered Python and never looked back. However, I revisit matlab about every 12 months or so for project purposes. I am really impressed with how far it has come. Despite it being very fast for some computations, I find the language cumbersome when constructing and end-to-end code to deploy on a cluster or a demo.

My vote: python

I've had a love-hate relationship with matlab over the years. I really like the symbolic package. But I don't need simulink for anything. Octave covers most of my needs.

When I was reteaching myself calculus it was with octave and matlab, it was funny that there were some integrals that octave could solve and matlab went into an infinite loop with, and vice versa.

To be honest, matlab made a lot of sense and was easier.

Everything became more verbose and harder to read when we switched to Python. Plot became plt.plot, a matrix A = [1,2;3, 4] became A = np.array([[1,2],[3,4]]), A=2:0.1:10 became linspace(,,,,), etc.

The object oriented programming doesn’t fit numerical calculations well.

Prototyping code is different from the production code, and doesn’t need namespaces etc.

I pirated Matlab just on a whim a couple years ago maybe like around 2019 or something and I never used it before in my life. I'm also not a programmer.

I found the built-in programming language in it amazing. The fact that I could type stuff out and then highlight it and then right click on it and then go to a very comprehensive documentation of the programming thing that I was typing that would also have an explanation for it along with multiple versions of code examples was a big help in being able to develop little useful programs for myself using it. I ended up using it to do batch processing of images or to find image quality metrics to compare different image compression methods. https://github.com/navjack/imagequalitycompare_matlab

It was all just built into this gigantic program.

I could even take the live code and eventually I figured out how to transfer that into the thing that allows Matlab to compile exes and make full GUI programs.

Like, I was able to build this with just context sensitive help in matlab's environment https://github.com/navjack/imagequalitycompare_matlab/blob/m...

I probably could have figured out how to make tons of useful things using it if for whatever reason my pirated copy stopped working or I had to reinstall Windows or whatever it was that made me drift away from using it.

I wish python had something that was like Matlab. I wish a lot of programming languages had something that was like Matlab. They just need to do exactly what Matlab is doing for in program context sensitive help that actually gives help. There is something just so special about the way Matlab did the help that wasn't dumbed down but it also wasn't too advanced. Like the documentation for each thing was extremely verbose but it was the right amount in contrast to other things where they don't explain anything or the code example is written extremely poorly.

But don't get me wrong. Later on I was able to make a couple very useful python things with the help of co-pilot. https://github.com/navjack/RandomGamePicker or https://github.com/navjack/ConsoleJournal or toy stuff like https://github.com/navjack/TextArtThing . I think co-pilot and running a local instance of fauxpilot is great but still not as fully useful as what Matlab does.

These days when I need to sketch out something in code for myself I tend to do it in powershell or python. I just need to get comfortable with a solid way to turn those ideas into a GUI program.

I guess I'll throw a wrench into your question and also throw GO into the ring cuz I think golang is neat. https://github.com/navjack/Go-Rainbow-Hearts-Term