[Grace-core] Should Grace abandon "gradual typing" for "like typing"?

Vitek, Jan j.vitek at northeastern.edu
Wed Jul 27 08:12:40 PDT 2016


Cool - that’s exactly what Thorn set out to do, and, I think, did.

> On Jul 27, 2016, at 5:10 PM, Andrew P. Black <black at cs.pdx.edu> wrote:
> 
> 
>> On 27 Jul 2016, at 7:05, James Noble <kjx at ecs.vuw.ac.nz> wrote:
>> 
>> I think I'm proposing that we don't investigate any of the more fancy sound, blame-carrying, etc gradual typing proposals and rather just stick with the
>> existing semantcs --- and point people to Thorn if anyone complains. 
>> 
>> Declare victory and move on, in other words - in particular to seeing how we can ensure adding more types doesn't make a program slower.
> 
> To me, those two things seem to contradict each other.   
> 
> My take on gradual typing — coming form the dynamic typing end rather than the static typing end — is that it’s all about figuring out how to ensure that adding more types does not slow down the program.   I think that Jeremy would say this differently — he’s looking at how to introduce the minimal number of type checks when adding type Unknown to  an otherwise statically-typed program.   But I think that it amounts to the same thing.
> 
> 	Andrew
> 




More information about the Grace-core mailing list