I think that it helps a lot to have a daily practice of using a language for small things.
In much that same way that many people do the daily wordle or crossword, I do the daily leetcode.
I flip a coin and solve it first in either C++ or Python, then re-write my solution in the other one.
Usually it takes me around 20 minutes to solve it in either language, and 5 minutes to re-solve it in either language.
Recently I decided to start learning emacs lisp. This is an imperative lisp dialect that’s pretty different from scheme, but I think that the particular language doesn’t matter much for this process. I could a bit biased because I do have prior experience with SML and scheme.
I started re-solving the problems a third time in emacs lisp. And I’m still learning but I’ve felt my comfort with the language increase over time, and I expect that if I continue doing this then I will eventually reach parity with C++ and Python.
Currently it takes me about 20 minutes to re-solve a problem in emacs lisp, because I usually have to read documentation and/or look up something new.
scheme is great, but the dx of some implementations is not. i’m on guile scheme due to guix, and frankly, i’m hating it a quite bit.
stack traces are esoteric and error messages entirely unhelpful,
documentation masquerades as deep but is indeed inconsistent and prosaic, mixing styles of reference, explanation, and how-to willy-nilly, (compare with Dybvig’s The Scheme Programming Language, which is focused and consistent, and it takes no time to get your answers; there’s just no method to guile and guix manuals), i hate it big time,
there’s big gaps in documentation (especially with Guix – there’s literally zero information about `define-record-type*` which is used everywhere in its codebase; admittedly, not scheme related, but still),
the cli requires too much memorization,
most modules are not named, but numbered, ie, instead of something like `(base list)`, you get `(srfi srfi-1)`, so you need to either memorize, or go through the info pages for each procedure you need to import, which means you also need to know the exact names for the procedures you need beforehand,
there’s like 4 ways to define a record, each with a different feature set and incompatibilities,
etc.
to respond to the content of the article, the different neurotype idea is off, because scheme allows you very well to express sequences of operations; the ecosystem of APIs may not cater to this tho. although if it was rephrased into “scheme emphasizes symbolic transformations, as opposed to machine operations” and how it does manipulate very different objects, i would agree
You can write entirely imperative OO code in Scheme, it just has syntax that's weirder than you're used to. Not everything has to be functional abstractions, syntax macros, or twisty mazes of call/cc. If SICP is too abstruse, give HTDP a try, but if you know other languages, you already know most of scheme.
Scheme was invented as a consequence of Sussman & Steele’s discovery that lexical closures in the lambda calculus had essentially an identical implementation to a fully elaborated version of Hewitt’s actor model.
I do wonder what a language with the same “taste” and minimalism as Scheme but embracing the actor model would look like. Erlang?
Even better if someone could figure out how to harmonize them in the same language: “There are exactly two ways to do it, and they’re interchangeable.”
There's something very ironic about an article about bouncing off Scheme on a website called 'SICPers'. OT: I think I'm pretty decent at thinking in Scheme, although I don't quite have the hang of continuations. That said, because I like type declarations I use Common Lisp, which allows me to bounce between a more Scheme-like style and a more Assembly-like style however I see fit.
I understand the challenge, but is Graham (OP?) getting too caught up in how the code ought to look, rather that what it ought to do. I don't think it matters much initially how a piece of work looks as long as it does what's intended. Afterwards it does; particularly if you need to involve other developers, and to them, the idioms looks "strange". I'm not convinced that there's an ALGOL neurotype that's distinct from a LISP(?) neurotype. I think it's a bit of a spectrum like everything else.
Strange that the post makes no mention of the Little Schemer series, because teaching you to "think Scheme" is exactly what those books do. Some people are put off by the weird style (combination of children's book visuals and socratic logic problem presentation), but they work!
I think it's the symptom of inadequate practice rather than some "language neurotype". Consider writing (yeah 2026 I know) a substantial project in Scheme from scratch.
I was the same way (and still am somewhat, I can't get hygenic macros into my head) but due to the differences between Scheme and Common Lisp. What helped me was writing imperative code that Scheme people would surely scoff at, and gradually using more and more Scheme features as I kept writing. Then I refactored the whole codebase to look like the final few hundred lines.
Programming languages, like natural languages, are tools for human beings, not computers. They work around the strengths and weaknesses of a human brain.
It's not a question of being smart or stupid. It's whether the tool fits the task it's applied to and the affordances it gives the user.
Scheme is intended more as a teaching tool than an actual language. Its simplicity is perfect for reasoning about programs. It's less well suited to practical tasks.
About the only really difficult lesson of Scheme is if you use it as a purely declarative language. Imperative features are a natural affordance of the human brain. Working with them is beautiful and alien.
In much that same way that many people do the daily wordle or crossword, I do the daily leetcode.
I flip a coin and solve it first in either C++ or Python, then re-write my solution in the other one.
Usually it takes me around 20 minutes to solve it in either language, and 5 minutes to re-solve it in either language.
Recently I decided to start learning emacs lisp. This is an imperative lisp dialect that’s pretty different from scheme, but I think that the particular language doesn’t matter much for this process. I could a bit biased because I do have prior experience with SML and scheme.
I started re-solving the problems a third time in emacs lisp. And I’m still learning but I’ve felt my comfort with the language increase over time, and I expect that if I continue doing this then I will eventually reach parity with C++ and Python.
Currently it takes me about 20 minutes to re-solve a problem in emacs lisp, because I usually have to read documentation and/or look up something new.
stack traces are esoteric and error messages entirely unhelpful,
documentation masquerades as deep but is indeed inconsistent and prosaic, mixing styles of reference, explanation, and how-to willy-nilly, (compare with Dybvig’s The Scheme Programming Language, which is focused and consistent, and it takes no time to get your answers; there’s just no method to guile and guix manuals), i hate it big time,
there’s big gaps in documentation (especially with Guix – there’s literally zero information about `define-record-type*` which is used everywhere in its codebase; admittedly, not scheme related, but still),
the cli requires too much memorization,
most modules are not named, but numbered, ie, instead of something like `(base list)`, you get `(srfi srfi-1)`, so you need to either memorize, or go through the info pages for each procedure you need to import, which means you also need to know the exact names for the procedures you need beforehand,
there’s like 4 ways to define a record, each with a different feature set and incompatibilities,
etc.
to respond to the content of the article, the different neurotype idea is off, because scheme allows you very well to express sequences of operations; the ecosystem of APIs may not cater to this tho. although if it was rephrased into “scheme emphasizes symbolic transformations, as opposed to machine operations” and how it does manipulate very different objects, i would agree
Official docs: https://docs.racket-lang.org/rhombus/index.html
Collection of small examples: https://github.com/racket/rhombus/blob/master/demo.rhm
I do wonder what a language with the same “taste” and minimalism as Scheme but embracing the actor model would look like. Erlang?
Even better if someone could figure out how to harmonize them in the same language: “There are exactly two ways to do it, and they’re interchangeable.”
There is Spritely Goblins: https://spritely.institute/goblins/
It's not a question of being smart or stupid. It's whether the tool fits the task it's applied to and the affordances it gives the user.
Scheme is intended more as a teaching tool than an actual language. Its simplicity is perfect for reasoning about programs. It's less well suited to practical tasks.
About the only really difficult lesson of Scheme is if you use it as a purely declarative language. Imperative features are a natural affordance of the human brain. Working with them is beautiful and alien.