John Cowan wrote: > Thomas Lord scripsit: > > >> Does the CL spec *forbid* adding those features? >> > > It has a notion of "implementation-defined attributes". >
So, the answer is "no," thus CL did not "regress" in the way that you claim. Since you recognize that that *would* have been a regression then, to be consistent, you should either agree that the restrictions in the R6 core are a regression or else convincingly distinguish the two cases. > However, you need no such permission from the spec to extend > the attributes of characters. If you want to give symbols > p-lists, you can do that too. > > I'm not sure why you bring that up, but... I've begun to think that, in the core, all values should (eqv-based) properties. MIT Scheme tried to do this by adding weak references and built-in hash values but a GC implementation seems to me to really have to stand on its head to get from those lower level features back to fast properties. -t _______________________________________________ r6rs-discuss mailing list [email protected] http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
