On Mon, Apr 27, 2009 at 11:06 PM, Grant Rettke <[email protected]> wrote: > > If you compare the Dybvig-Ghuloum model and the PLT model, when is one > "better" than the other?
When you have no need to import identifiers at a specific level only. Notice at the moment I do not know of any convincing use case where one would like to import identifiers to a specific level only, but I could change my mind if you have one. P.S. the multiple instantiation semantics of PLT is not a big deal, since Ikarus can emulate it by using separate compilation: by construction then module instantiations would be separated, since they would happen in difference OS level processes.
