The code itself is here: https://dfns.dyalog.com/tube_code.htm
I'm not an expert in APL-family languages, but they seem to be perfect for those who have developed an intense hatred of boilerplate and explicit loops.
Array languages in general, note that NumPy traces back to array languages.
Also that is a great reason why array languages would be ideal for GPGPU, and sadly we're stuck in more classical languages with extensions for that.
Three pages, including comments. Always amazing to see that.
Sometimes I think I understand computation, then I look at APL.