I am on vacation. This newsletter is only partially for you and very much for me. So it goes on while everything else takes a month-long breath.

Considered Exploration of the Mainstream

Sometimes I stick with Elixir because it fits my particular brain. Sometimes because it is the best tool for the job. Sometimes because it offers enough good options that I don't have to tackle things I am not very interested in. I distrust that last one.

Overall, I may prefer my particular tools but it is usually a mistake to not have a decent understanding of the mainstream offering.

I try to be particular in my criticisms of JavaScript. I don't dislike JS. I kind of love the quirky weird language and all it let me do since the first day I used it to make an animated menu with images on hover. I typically say that I don't enjoy the JavaScript ecosystem. I've said it so many times that the state of things has probably mutated under my feet. Granted, I am obligated to use parts of it often enough still that I do know it to some extent. But I am not immersed, I'm not trying to grow my JavaScript ecosystem skillset.

TypeScript is a thing I've avoided like the plague. This is probably dumb. I just listended to Oxide & Friends and that optimistic view of TypeScript and the tooling they used gave me a glimmer of hope that it doesn't suck. Of course it is a rather optimistic podcast so I'll take it with a grain of salt.

I have never liked static typing and have never wanted it for myself. So everything about TS seemed like JS but more overbearing. "Oh great, C# again!" was my feeling. Nevermind that my problems with C# is not actually type signatures but rather enterprise OOP.

A two year bath in Elm has shifted my tolerance for static typing and added a smidge of appreciation. I respect the model of Elm and Rust where you try to design away entire classes of problems. It is not the fastest way to a result but it may be the one that requires the least rework and builds the most confidence. And if it is, it probably does pay off.

If TypeScript is this but more pragmatic about interactions with the browser (one of Elm's sharper corners) then it is probably worth figuring out if I do hate it or not. And I guess I get it "for free" with Esbuild.

I am also considering asking a knowledgeable React-dev I know to just show me how he does it. Make React look good to me. I want him to steel-man React + TS for me. Because I have so many layers of skepticism towards the stuff and some of it is just cultural tribalism (which I don't really value) and some of it is exhaustion (having heard so much about it and not being interested). Some of it is my small and frustrating experiences with it.

React doesn't seem like the right move for me so I doubt I'll like it but I wouldn't be surprised if TypeScript is actually fine. And then I have another trade-off to make in the future. Do I want a build-step or nah.

So that's frontend. I have no interest in JS/TS for the backend.

If I were to approach the mainstream on the backend, Rust would be my first choice. This is not driven by Oxide & Friends though they've probably also encouraged me in that regard. Rather I see a lot of interesting development over there and I've worked through the free official book thing a fair bit and the language makes sense to me. It would be a new useful tool in my kit. Though Isaac told me in this video that I just need 3 days for Zig.

I don't think revisiting C# or picking up Go would give me much in terms of new capabilities. I'd happily lear to do Go if I had reason to though. I like the static binaries.

In ops it seems I should spend some time with Kubernetes and much like with JavaScript/TypeScript see if my current set of reservations hold up. It seems a complicated beast and unpacking that to understand it is bound to be good learning.

The thing is, I don't see myself "moving on" from Elixir. Not now and probably not for a long time. I've seen enough people do Erlang until retirement age that it seems like a perfectly valid end state to build on the BEAM. I still want to learn though and there are things that probably shouldn't be built on the BEAM (local-first web, arguably native desktop apps, mobile apps). I like being opinionated about tech. It is satisfying to say no to things and opt for simplicity. That doesn't mean I should avoid understanding the landscape.

What are some things you are avoiding learning that you should probably pick up? You can reply to this email to reach me or poke me on the fedi @lawik@fosstodon.org.

Thanks for reading. I appreciate your attention.