On Mon, 4 May 1998 15:50:35 -0700,
Barry Margolin <······@bbnplanet.com> wrote:
>Sometime around 1991 or so, I think there was an article in SIGPLAN Notices
>or LISP Pointers that evaluated CLOS with respect to a well-known set of
>criteria for OO language features (I think they were from OO bigwig like
>Booch). Except for the information hiding features (which the article
>admitted could be done somewhat with packages), it scored reasonably well.
I have only now managed to find the exact reference for another
source that might be interesting. The September, 1991 issue of
Communications of the ACM (Vol. 34 #9) has a number of articles
devoted to Lisp. Without any implication that the rest are
uninteresting, in the first one entitled `CLOS: Integrating
Object-Oriented and Functional Programming' (pp. 29-38)
Gabriel, White and Bobrow `describe the components of CLOS
and show how CLOS related to the object-oriented languages
Smalltalk and C++.'
I shall quote just this remark relevant to the issue of
hiding slots (p. 36):
However, CLOS has no linguistic support for mixins, as Flavors
does [13]. Such support would constrain the use of classes
defined as mixins, or provide a means to hide names in a class
so that, for example, slot names in mixin classes would not
conflict with the slot names in the classes with which they
are mixed.
(Reference [13] is Symbolics, Inc. _Symbolics_Common_Lisp---
Language_Concepts_, Chapter 19 ``Flavors,'' Symbolics, Inc.,
Burlington Mass., 1988.)
By the way, the _Art_of_the_Metaobject_Protocol_ devotes
some space on the above issue, too. I don't have the book
at hand right now to give an exact reference, but it's not
hard to find (Chapter 4, I think).
An additional question: does anybody know if the picture on
the front cover of AMOP depicts actual people (perhaps the
authors of the book)?
Best regards,
Vassil.