NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/lang/clos/1003

Internet Message Format | 1992-11-17 | 1.7 KB

Xref: sparky comp.lang.clos:1003 comp.lang.lisp:2893 comp.lang.eiffel:1310
Path: sparky!uunet!ferkel.ucsb.edu!taco!rock!stanford.edu!ames!saimiri.primate.wisc.edu!zaphod.mps.ohio-state.edu!usc!rpi!ghost.dsi.unimi.it!univ-lyon1.fr!chx400!sicsun!disuns2!/!matomira
From: [e-mail] (Fernando Mato Mira)
Newsgroups: comp.lang.clos,comp.lang.lisp,comp.lang.eiffel
Subject: Type declarations and generic classes in CLOS - HOW??
Summary: Here we go again...
Message-ID: <[e-mail]>
Date: 16 Nov 92 17:25:18 GMT
Sender: [e-mail]
Reply-To: [e-mail] (Fernando Mato Mira)
Followup-To: comp.lang.clos
Organization: Ecole Polytechnique Federale de Lausanne
Lines: 30
Nntp-Posting-Host: disun47.epfl.ch

Let's say you have a class:
(defclass foo () ((x :accessor x :type number)))


and you would like to be able to do something like:
(proclaim (type (foo integer) an-instance))


so a smart LISP compiler can avoid type checking when you write:
(proclaim (type integer n))
(setf n (x an-instance))

How would you go about? It seems that generic classes (and the general issue of
type declarations - see article "Declaring types by association in CLOS")
have been overlooked in the design of CLOS.

I am trying to do VR, so every declaration counts.

I remember a paper by Luca Cardelli on an implementation of ML with a
higher-order type system. Shouldn't LISP have something like that to
stay at the cutting edge?

--
Fernando D. Mato Mira
Computer Graphics Lab               "I must admit it, I have learned
Swiss Federal Institute of Technology            some things from Eiffel"
[e-mail]

NeXTMail : [e-mail]
FAX      : [phone] - 5328