I'm debating with Shap whether or not it's OK for type inference to
sometimes fail for unintelligible reasons. I say yes, Shap says no. My
motivation is that getting well-behaved type inference places
uncomfortable restrictions on the language design, at least given the
current state of the art.

At least that's what I think we're debating. So I don't think anyone
has a problem with allowing type annotations wherever you want.

On Thu, Jul 10, 2014 at 5:02 PM, Raoul Duke <[email protected]> wrote:
>>> But the more important issue is that humans can't write static types
>>> correctly. In today's systems, the types involved are complex enough that
>>> when humans are asked to state them they come up with typings that are not
>>> complete.
>>
>> Are you saying _no one_ will understand the type system well enough to
>> write good types? I don't think so. It sounds like you want BitC to be
>> useful to people who don't know how to use it.
>
> i don't understand the point (...either; or possibly just me). for
> typed languages with inference i don't think i've ever used one that
> outright prevented me from annotating with my own types if i want? and
> i am under the impression that furthermore inferencers always go wrong
> somewhere so the system much allow humans to interfere.
>
> even just looking at it the other way around: what if i want to write
> my types first and then fill in the details. i think that's a valid
> thing and should be supported. or, write code then write types and
> then fix my code to fit the types (i've done this, it is similar to
> writing the unit test and then fixing the code to just pass the test;
> it is a weird-good experience when it works).
>
> i suspect this is all wooshing way over my head, apologies.
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to