On Fri, Jul 4, 2014 at 3:21 AM, Gabriel Dos Reis <
[email protected]> wrote:

>
> One thing I've found is that programmers (as opposed to type theoretists)
> don't really care about strong normalization.  They care that the
> computation they are doing at compile time is terminating (or not)....
>

I tend to agree. The only reason that I am tentatively holding on to strong
normalization is that it induces a simple theory of type equality. Though
I'm not entirely convinced that we will turn out to need that. I suspect
that in our use cases it will turn out that the compile-time operations on
types will turn out to be limited to construction and
destruction/inspection. That doesn't guarantee strong normalization, but I
suspect that normalized results will tend to emerge from the [human] use of
these two operations. I also suspect that concerns about equality will end
up to be sufficiently covered by unifiability and subsumption.

We'll see, I guess. :-)


shap
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to