Now that I've done posting about why I'm retiring we can get back to our usual blogging.

I haven't really done anything technical over the past couple of year. The most programming I did in the past 18 months was a halfhearted effort at the Advent of Code 2021 event last December. I thought I'd see about picking up a new language to change that. In doing so, I was reminded of some of my resistances to learning a new language and how I've seen similar from self taught students coming into my classes.

Of course my situation is different - I'm doing this wholly by choice and once a student is in a class they've got to use the prescribed tool.

The contenders on the functional side were Haskell and OCaml and on the other side, mostly Rust.

Today, we'll talk about the functional side. I decided to look at OCaml. I had briefly looked at it well over a decade ago but barely scratched the surface. Haskell I played with a little more deeply but not much. I decided on OCaml because it was supposed to be more real world and, to be honest, I get turned off by some of the holier than though purely functional talk that sometimes occurs.

Funny thing is that I've had a really hard time getting going - so much so that I decided that for now I'd explore Rust.

So, what's going on?

OCaml looks to have some nice features. It's functional and I'm a fan there and it's also supposed to be pragmatic - you can get real world stuff done with it. It's also statically typed which overall I think is a good thing and it has cool features like pattern matching.

For pattern matching, I'm not talking about things like regex or matching characters in a string but rather something like a really cool and powerful case or switch statement. Even that description, though is not doing it justice.

I think my problem is that I already am comfortable and enjoy programming in Clojure. Clojure is functional and pragmatic but while Clojure doesn't have pattern matching built in but there is a library that basically accomplishes the same thing. As to the type system, Clojure is dynamically typed and while that would be an issue if I were working on large real projects and/or on a team, I'm just writing little toys by myself. If I want stronger typing, there are ways of doing that in Clojure. True, it's not required but working by myself, I have that option, should I choose. It's not the same - I know but as I said, I'm not doing anything big, important or with other people right now.

So, basically, every time I try to dive in to OCaml, I end up feeling that "I can already do that and do it better in Clojure." I'm guessing I'd feel the reverse had I started on the OCaml side.

This has happened to me before. I was an early Python adopter feeling that it was a huge improvement over Perl. A while later, Ruby came out and shortly after that, Ruby on Rails became a thing. Every time I looked into Ruby I had to stop - I could already do and do better in Python. By then, there were enough tools to compete with Rails even though they weren't as mainstream and overall I preferred Python. I never did end up learning Ruby.

I felt a similar sluggishness when I went from C++ to Java - something I had to do because of the AP Exam change but I pushed through since I had to teach it and ultimately, I've learned to appreciate the good and curse the bad of both languages.

I've seen similar blocks or resistance from who've had prior programming experiences over the years. From students who seemed to be taught more "the language" rather than CS using the language. I know that's a nebulous thing to say but I think some of you will catch my drift. It's subtleties like saying "okay, we're repeating this process, here's how we do it in language such and such" vs "here's how we repeat this process period" and show the language construct.

My former math chairman at Stuy said he once, back in the 80s asked Marvin Minsky what they should be teaching for CS in High School. Minsky said, and I paraphrase, "nothing, you'll only mess them up. At most teach them to type." Of course that was a different time and I strongly disagree with that statement but I do think we have to look at how students entering our classes are or are not prepared and adjust accordingly.

So, in the end, it's back to Clojure for me on the functional side and I'll try to do an OCaml dive sometime in the future.

Fortunately, I'm having an easier time getting in to Rust. More about that next time.