Rust development log: day 100. Or so. It’s starting to get fuzzy.
Programming is hard. Learning new things in programming is especially hard. When all you know is imperative Python and you move onto declarative SQL, expect to have to wrap your mind around a lot of new challenging ideas. Don’t expect it to be easy.
Beginning anything fundamentally new feels like staring at a brick wall that’s blocking your path. The wall is impassable until a point that you make a key discovery which lets you make your way through. Often that discovery is a eureka moment – it felt impossible only hours before.
But eventually you find your way into the clear. The walls get smaller and you navigate passed them more easily. At some point you might gain so much mechanical dexterity that you find yourself in an open field – you know the language and its tooling so well that there’s practically nothing that can stand in your way.
My experience so far with Rust is that this doesn’t happen. In the beginning, the walls are higher than in other languages, but surmountable: learning the ownership system, understanding references, building intuition around lifetimes.
Eventually you become a more experienced Rust programmer, but things don’t get much easier. There’s always another wall waiting around the next corner. The horribly anti-user futures system, compiler messages generated by a misused macros that are second only to C++ template errors in how egregiously difficult they are to parse, universally terrible documentation and lacking examples, unstable APIs, type annotation hell, and so much more.
I’m still undecided on the language. I love that Haskell-esque feeling where a compiling program is usually a working program. I like the syntax and the tooling. I don’t like how little time I spend solving the problems that are interesting to me. I estimate that I spend about 5% of my time building new domain logic, and 95% of my time fighting the compiler over problems that are appalling in how uninteresting they are. Even something as simple as abstracting a new helper function often turns into a veritable odyssey because getting type annotations right can be so difficult (especially where a third party library is involved).
These thoughts are not intended to be comprehensive, and I’ll write a more substantial defense at some point, but I wanted to sketch a few feelings first. I intend to keep pushing forward, but if forward progress doesn’t get easier in a few more months, I’ll probably give it up (hopefully with a post-mortem). Learning is always valuable, but I’m starting to feel seriously concerned by how glacially slow it is to get anything done.
Did I make a mistake? Please consider sending a pull request.