/tech/ - Technology

People would have these awesome programs that were written by this mysterious programming language that the author kept promising to release, but then he died and all the attempts to reconstruct it just end up with overhyped boring plain old programming languages.

>>21689
So… HolyC?

Lisp but purely event-driven, with portable call stack reflection and implementing the negate and sublate functions, which work as you would expect.

>>21684
computer science is a formal science so it uses mathematical logic

Like Verilog, electronics is the thesis, Verilog the antithesis and hardware the material synthesis.

>>21692
The separation between mathematical logic and real world systems is a big problem in software development, or at least used to be a big problem. Nowadays it's not so much in most systems, and you can happily program in very high abstractions without worrying about pesky hardware issues.

That said, when we do software, we do it to solve a real world task. The biggest pain in software engineering is transferring the complexities of the development of a real world task, into a relatively static representation that can solve the task. Big issues arise when the real world task changes or more context is introduced. Then the old digital representation is no longer valid and needs to change. Unfortunately, software becomes "hard" the more you change it and the bigger it grows.

If there is to be a programming language that has some passing resemblance to marxist philosophy, then I suppose one that can simultaneously capture the complexities of the domain to a large degree, while not necessarily delivering on the implementation of the domain.

Meaning that you could write a program incrementally by implementing a highly detailed domain description. This way, the language itself could coerce you into writing implementations that actually fit the domain, without any incorrect assumptions since they're already spec'ed elsewhere.

For this to work, you'd need to have all the "stakeholders" be capable of writing down all the assumptions and constraints of the system, including where they can vary, and where it's absolutely impossible for them to vary.

Afterwards, if conditions change, then the specs need to be changed first, which would cause the implementation to fail to build, which would prompt a re-implementation or a fix.

Just riffing about it here. I don't see this as being particularly viable.

HASKAL

>>22020
You mean haskell with wirthian syntax?
I tried this with the example from haskell.org:

function FilterPrime(List PrimeList) return List is

  function ListPredicate(Integer x) returns boolean is
  begin
    return x mod List.head /= 0;
  end ListPredicate;

begin
  List(PrimeList.Head, FilterPrime(List(PrimeList.Tail, ListPredicate))));
end FilterPrime;

List Primes := FilterPrime(2 .. Integer'Last);

We'll only know when communism happens.

>>21691
>and implementing the negate and sublate functions, which work as you would expect.
????????

>>21995
>Meaning that you could write a program incrementally by implementing a highly detailed domain description. This way, the language itself could coerce you into writing implementations that actually fit the domain, without any incorrect assumptions since they're already spec'ed elsewhere.

3:18 onwards

>>21684
1. as a prerequisite you must read ten 1,000 page manuals
2. each manual contains long drawn out histories of how the authors of the previous manual were traitors to the project of creating a new programming language, and contradicts the recounting of the previous authors
3. instead of printing stacktraces it accuses you of being schismatic