[Haskell-cafe] WANTED: Compensated Haskell Hacker for Language Project

[Bryan Edds]

Hi Jake!

> My only question is this: what does your language offer that others do 
> not with respect to soft real time systems? The language you describe in 
> the linked forum thread looks neat, but I think I'm missing the 
> reasoning behind its design. Why is this design beneficial for soft real 
> time compared to other high level languages?

The main thrust of the design is to provide nearly the power of Lisp and ML's
semantics in a form that is syntactically palatable to the mass of intelligent
industry programmers. While industry programmers typically prefer C-style
languages, it's just not possible to build a C-style language with a reasonable
macro development (language orientation) experience due to C's inherent
syntactic complexities. Further, it seems to have been historically demonstrated
that C-family programmers are not willing to make the a syntactic leap as far as
say, Lisp or Ocaml.

Barring the provision of yet another C-style language, there's another set of
languages many C-family programmers do rather like: Ruby and Python. So by
finding a direct mapping from s-expressions to a language with an feel and
visual appeal similar to Python that ALSO approaches the machine efficiency of
C++, I hope to create a lisp- and ML-derived language that is accessible to an
audience wider than existing functional languages seem to have reached.

As you can see, the design does admit some semantic compromises in the name of
syntax and efficiency, but the compromise is surprisingly (at least to me)
minimal. One compromise made in the name of C++ efficiency is the use of a
machine word-sized default number type rather than the default number type used
in lisp or Haskell. Of course, arbitrary number types can be made available
naturally via a library using simple binary operation overrides, but they are
not the default when you type the literal 5.

I've discovered, at least to my own current satisfaction, that bringing lisp-
and ML-style semantics to the masses is possible with less semantic compromises
than previously seemed possible. By providing a language that industry
programmers feel comfortable with, I hope to move as many of them as possible
into the functional / language-oriented world.

But more than that, I'm just designing the language I wish I could use everyday
instead of C++ and C# :)