Exactly. There's an art to using the right tool for the right job. C++ has some good use cases. Mathematica has some other. Excel do excel in many ways of prototyping, trying out simple but flexible things, and learning fast.
One thing is the process of learning it. Then you can spend time playing with it. Just have fun!
Something else is to design something that will bring some value to you. What kind of qualities are you looking for, for which purposes? Something for research? Trying out creative / innovative ideas? Heavy-duty calculation? Real-time execution? Here also, it should be fun! And not too much work or complexity to maintain. The idea that one size should fit all, for simplicity, is not a practical ideal to aim for.
Yes, it's *possible* to do all that in C++ or even Excel. Experience is knowing that the end result doesn't always have to be 100% correct, too structured / rigid or even beautiful. Often, there's value in making something in a few hours just to try it out, and avoid overoptimizing it too early, which usually become a big waste on the average. If you tell yourself beforehand you will throw away all your code, and do it, you become so much more creative and efficient. At least don't marry your code. It's like marrying a position.
I've lost count over the years of time spent coding on useless things that never became truly useful. I mean, I learnt alot, but optimizing and grinding out libraries, was just useless as an end result by itself.
