-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Dave Berry wrote:
Hi Lucas,
It's more complicated than that. If you want to pursue this, you need to get hold of the formal semantics of Standard ML and construct proposals in terms of that semantics. There are people in Informatics who should have copies of the definition and the "commentary" that accompanies it.
Hi Dave,
I am not sure it is more complicated than that: structure sharing, as of ML 97, is an abbreviation for sharing the types in the structure (sharing type ... = ...) and given David Matthews statement that this is only allowed for unconstrained type identifiers, I'm suggesting the abbreviation is only expanded for unconstrained type identifiers.
This doesn't affect the semantics as I'm just suggesting a different abbreviation that happens outside the semantics. Or is there something else I missed?
lucas