Re: ISSUE-91

Too much sugar leads to bloat and rot.  It's not very good for language 
design either :-)

This has been Christian's pet PRD feature for so long, I've grown tired 
of arguing about it.

I think all the dialects, PRD included, can do without this sugar.  I 
prefer treating all the conditions uniformly, i.e.
Forall (x) Q :- P AND C(x)

Dave Reynolds wrote:
> Chris Welty wrote:
>>
>> At the F2F12, there was discussion of "bounded quantifiers" (e.g. 
>> Forall (x in C) ...) in PRD, and the suggestion was made to put them 
>> in Core, and thus into BLD as well.
>>
>> This would require re-issuing LC for BLD.
>>
>>
>> It seems to me this could be done simply as syntactic sugar, ie
>>
>> Forall (x in C) Q :- P
>>
>> is syntactic sugar for
>>
>> Forall (x) Q :- P AND C(x)
>>
>> and just allows an implementor to more easily recognize the 
>> restriction on the quantification (this is a common source of 
>> optimization in implementations).
>>
>> Anyway, let's have a brief discussion about the pros/cons on Tuesday.
>
> It seems like a simple bit of syntactic sugar and if that noticeably 
> helps uptake then it may be worth doing. I certainly wouldn't object 
> to it if that is the consensus.
>
> However, I don't quite see why detecting that there are terms of the 
> form C(x) in a rule premise is such a hard job and so why a custom 
> syntax really helps that much with implementing such optimizations.
>
> Would the syntactic sugar really be needed in Core (and thus BLD) or 
> would it only be needed in PRD?
>
> Dave

Received on Monday, 16 March 2009 17:38:15 UTC