2010/3/11 Jonathan S. Shapiro <[email protected]>:
> Philipp indirectly raised a valid question: given the goals of BitC, does it
> make sense to be trying to do bring-up on CLI. I have reservations, but I
> think the answer is yes.

It does make sense to bring a safe language to a popular and safe
platform with a good garbage collector. And because Java still doesn't
have proper tail calls, CLI is a natural choice between the two. I do
believe that modifying BitC's type system to understand the full CTS
is an anti-goal, so BitC could never be a full-fledged .NET language:
you can't use .Net interfaces without full support for the type
system.

>From the marketing point of view, CLI isn't really that good:
 - you cannot do smooth interop with .NET
 - the performance suffers due to non-optimal runtime
 - CLI already has languages such as F#, Ada/Spark, Spec#, C++/CLI and
static analysis tools such as Code Contracts (which is bundled with
.NET 4)
 - replacing C with something safe isn't a good argument anymore

On the plus side, having a high-end garbage collector does help, even
if it's not optimal for functional languages. And there aren't many
open source compilers for CLI langauges.

To the point: I think CLI is an important target, but I do believe
that the native (or LLVM, C-- or C) backend should have the priority.
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to