Simplicity, maybe? I think it's a shame that Dr. Xi's work on DML hasn't had
more of an impact on other languages, but there's a cost to every feature,
and as you note, ATS has a lot of features.

I've only read a few hundred lines of ATS code, and written none, so my
opinion is worth horsefeathers, but the impression I got was that a rough
predictor of unreadability is the square of the number of type system
features used beyond basic ML. The example code for a function that provably
sorts its input list is cool in theory but (at least to my untrained eyes)
impenetrable in practice.


On Sat, Sep 4, 2010 at 7:56 AM, William Leslie <[email protected]
> wrote:

> I have been considering BitC's goals recently and wondering if they
> might be fulfilled by the language ATS.
>
> ATS has good compilers, a reasonably trendy syntax (similar to Agda),
> and the best C integration I've seen in a long time. It also has a
> fairly low-level view of data types and mutation, closures, as well as
> dependent types, linear types, and even an interactive proof system.
>
> What does BitC provide that ATS doesn't?
>
> --
> William Leslie
> _______________________________________________
> bitc-dev mailing list
> [email protected]
> http://www.coyotos.org/mailman/listinfo/bitc-dev
>
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to