What I like about "tests" in software development is that anyone can run them, just download the source code, then run ./test or right click and "run tests". It would be cool if computer science could offer the same experience, just download the source code and run it, compare if you got the same result, inspect and learn from the source code, etc. Instead of "here's some pseudo-code we've never tried", and here's a mathematical formula that you need to be a mathematics professor to understand... Yes we know you are not a professional software developer, the code is going to be at a beginners level, but that is fine, I am not reading your paper to criticize your code for being "impure", or not using the latest syntax and frameworks, I'm reading it to understand how to implement something, to solve a problem.
I also cannot understand how a paper whose main contribution is a set of benchmarks, does not actually make the source code to those benchmarks publicly available. Unbelievable that such a paper can pass peer review. Very unscientific.