Working Up

by Dwayne Phillips

We used to discuss this topic, but with different choices. Time passes.

People discuss this on the Internet. Today’s question is “R or Python?” Good grief, two interpreted languages that run slow. But they are good for prototyping things, bringing in mammoth libraries, and writing useful things in ten lines of code.

We used to discuss LISP or Prolog. In our lab, we were moving out FORTRAN and moving in C. C had better data structures while both had efficient execution. Being efficient was quite important when your minicomputer (a couple racks of equipment) had 128KiloBytes of memory. I wanted to use Ada, but we didn’t have any compilers on our machine.

Unix was coming in. Folks in industry were using DEC machines with VMS. University research labs were on the leading edge with one or more flavors of Unix. Linux would come ten or more years later.

It all worked quite well, and we made advances. Less computing power meant more brain power. That is not a brag from an old man, but a simple tradeoff. More computing power means more things are possible in less time. Why sweat details when you have the machine do the work? That is a tradeoff and a good trade.

What programming language for artificial intelligence? What problem are you working? What are the tradeoffs? Think. Choose. In that order.