You’d expect the longest and most costly phase in the life cycle of a software product to be the initial development of the system, when all those great features are first imagined and then created. In fact, the hardest part comes later, during the maintenance phase. That’s when programmers pay th
Depends what you mean by pure-functional. If you mean an ML-family language with lazy evaluation and explicit sequencing of side effects, then probably not. But there are certainly lisps suitable for those constraints. uLisp runs on Arduino Nano microcontrollers with 32 Kbytes of program memory and 2 Kbytes of RAM. lllm is a lispy DSL for writing assembly language.
Yeah, probably also worth pointing out that the original LISP came out in 1958.
A function isn’t complex to portray in memory when you’re passing it to other functions. It’s also just a pointer to where your function’s code lives. @ttmrichter@lemmy.ml
Lisp is “functional-enabling” not a functional language. Take a look at the Common Lisp library and you’ll see a whole lot of imperative-nature constructs and code in it. Lisp permits (and to a degree supports) functional programming. It does not enforce nor even default to the functional paradigm.
And even in a Lisp you’re going to start hitting the boundaries of your kit in the embedded world when you do partial application, closures, and even something as seemingly trivial as using immutable state. You’ll wind up fighting the language more often than fighting your problem domain and the result will be counterproductive.
Best of breed for me in embedded space is Ada (with Modula-3 being another decent choice). Lisp’s not even in the display hall.