Following a conversation with Scott Smith this morning, I'm dropping all
plans to implement existential types in BitC. Existential types are a
large, mathematically complex hammer. All of the practical use cases
that we can identify can be resolved using a simpler approach (objects)
that is closely related to something that we already have
(make-closure). I'll explain that in a moment.

shap

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

Reply via email to