Am Mittwoch, 22. März 2006 14:19 schrieb Bulat Ziganshin:
Hello Wolfgang,
you said WHAT you think but not said WHY? my motivation is to be able to use myriads of already implemented algorithms on new datatypes
I think, I already tried to explain why I think the way I think in an earlier mail:
Strictness has to refer to attributes (the things you apply a data constructor to). In you approach, strictness is connected to type arguments. This causes problems. For example, if you have
data T a = C a a,
what would T !a mean? Would both attributes be strict? But how would you force only one attribute to be strict then?
By the way, would it be okay for you to answer below the quotation, not above it? And would it be possible to use just a > sign, followed by a space for marking quotations. My MUA gets confused by things like “WJ> ”. Thank you very much.
[...]
Best wishes, Wolfgang