Over the past two months, I've written code in 50 different programming languages (with an eventual goal to get to 100). Here are my thoughts thus far.

Overview

I started this journey because I wanted to try new programming languages in a low-stakes setting. Rather than learn a language in depth, I'm just writing a page of code, to see what either the programming paradigm or development experience is like. Since I'm focused on new (to me) languages/paradigms, the majority of the languages aren't even in the top 100 programming languages in the TIOBE Programming Community index. I'm definitely not trying to improve my job prospects!

Positives

Overall, this experiment has been a success. Prior to starting, I thought I'd already used a wide variety of programming languages, but I'd never actually used an array-based language, and I didn't even know what concatenative languages were (thought I'd briefly heard of Forth)! I'd also never experienced column- or line-based input or written right-to-left text.

If you've got the time and desire novel programming experiences, this is a worthwhile journey!

Negatives

The last sentence implied it, but repeatedly starting over from scratch--often in new paradigms--takes time. I'd say that on average I spent at least two or three hours with each language--and often more. Frequently, it was worthwhile, but stumbling through Smalltalk-like UI and syntax felt more tedious than educational, for example.

It's not obvious, but I also spent a fair amount of time discovering, selecting, and occasionally shelving languages.

The road ahead

Additionally, now that I'm 50 languages in, the novelty is wearing off a bit. This is welcome, in that the learning process is less taxing, but I'll need to ensure this doesn't just become a slog to 100. If I stop learning and/or enjoying the process, I should quit. Fifty programming languages is no small feat, anyway--though I'll feel a bit silly having a repository named "100-languages" if I never quite get there.

Tangentially related: I feel like there is a distinct possibility that I will end up creating a new programming language either during or after this exercise. The world definitely needs another Lisp (or maybe Forth)!

Languages by category

For the record, here is a rough categorization of the languages I've used so far (with categories based mostly on this article). Note these classifications are subjective!

• Assembly: WebAssembly, x86 ("real"/16-bit), Uxntal
• Imperative: Fortran, COBOL, Perl, WebGPU Shading Language, Awk, Commodore BASIC, BBC BASIC, TCL, Bash, Turbo Pascal, Hare, wax, Nim
• Homoiconic: SectorLISP, fe, Rebol, Boron, Cakelisp
• Functional: Standard ML, Unison, Julia, Pkl, Roc, Haskell, Idris, F#
• Object-oriented: Squeak, Wren, Self
• Concatenative: Gforth, PostScript, RetroForth, Uxntal, Factor, Min
• Array-based: J, Lil, APL
• Declarative: SQL, XSLT, Verilog
• Block/flow-based: Scratch, XOD
• Esoteric: SIC-1 Assembly Language, EXA, Piet, قلب (Qalb), 文言 (wenyan)

Excluding intentionally esoteric ones, the most mind-bending for me have been concatenative and array-based languages, probably due to my inexperience with them. Dependently-typed languages would probably have been mind-bending if I'd spent more time and actually learned them.

Programming language observations

Being a newbie nearly 50 times in quick succession theoretically gives me a unique perspective on programming languages. Here are a few observations.

Ramping up

Overall, for me as an experienced programmer, the most helpful thing for ramping up has been, unsurprisingly, quality documentation. Specifically: having searchable documentation that is clearly organized, using familiar language (e.g. outlining basic syntax, enumerating control flow options, and diligently documenting the standard library).

The languages I struggled most with seemed to focus their documentation too much on tutorials and examples (usually unrelated to what I wanted to do--which is fair) or they used long-winded prose. In one case, the "hello world" example contained a URL that wasn't even explained--that just made me more confused, rather than less. Thankfully, none of the languages tried to rely on videos or chat apps for "documentation"!

No one cares about footprint

It's disheartening, but most programming languages don't seem to care about the disk/memory footprints of their environments or, often, their resulting binaries. I don't enjoy slow software, but, these days, my old computers are more impacted by bloat than sluggishness.

Hare and Rebol are nice counterexamples, and embeddable scripting languages are also usually pretty light. And let's not forget the old, but famously lightweight Turbo Pascal environment!

So, so many languages

Despite my frequent complaints, the quality of virtually all languages I've tried is impressive--even the hobby languages are eminently usable! The amount of work involved is staggering.

Perhaps people create new programming languages for the same reason they create art, i.e. because they can't not do it?

Parting thoughts

Now that I've sampled a lot of different programming languages, I'm more confident about matching problems to appropriate languages. In the recent past, I've tried to find the one true language that I can use for everything, but that's honestly misguided--most of the time, at least--and especially for hobby projects in wildly different domains.

I guess the good news is that I've now convinced myself that I could probably make do with any programming language, so maybe I can finally stop bike-shedding about programming languages.