re: generic-function, generic-flet, generic-labels

[Dick Gabriel <RPG@SAIL.Stanford.EDU>]

[In reply to message from KMP@STONY-BROOK.SCRC.Symbolics.COM sent Tue, 13 Feb 90 13:02 EST.]

I am the champion of them. I believe that for unimportant reasons their
syntax was misdesigned, but I think they are very important. I wish we had
a CLASS-LET as well. There is a tension between placing functionality in a
language for the purposes of achieving environmental goals and demanding
that the environment handle those goals. There is no need in CLOS for
DEFMETHOD except to be able to display methods near the classes to which
they apply. However, with multi-methods, one is forced to choose one of
those classes, and maybe not all of them. An environment should let you
look at these methods anyway you like. The real linguistic need is for
constructs to build generic functions, which is what those three
constructs do. Incremental redefinition can be handled by the environment
or by ADD-METHOD, FIND-METHOD, and REMOVE-METHOD

			-rpg-