I think it was made clear in the article that I used a naive implementation in all the languages. The point is: I can optimize the code in each language, but are they comparable then?

By standardizing the code I can compare how languages perform doing the same operations, rather than features available for optimizing execution time and memory.