🚀 go-pugleaf

> Lucid's implementation would then meet all your criteria if only
> CLOS::SLOTTED-CLASS had a direct element that was a direct superclass
> of STANDARD-OBJECT and FUNCALLABLE-STANDARD-OBJECT, no?
Yes, if we can agree on a name for this class, and a meaning for the
class STANDARD-OBJECT.

> This is
> a point outside the standardization rignt now,
Maybe so.  Are you referring only to the ANSI Common Lisp document,
or also to either CLtL2 or Chapters 5 and 6 of The Art of the 
Metaobject Protocol?


>> The classes STANDARD-OBJECT and FUNCALLABLE-STANDARD-OBJECT are meant to be
>> "maximal elements" of some set of classes.  The disagreement is over which
>> set.  There are really three sets of classes that are relevant:
>>   (1) The set of all subclasses of STANDARD-CLASS,
>>   (2) The set of all subclasses of FUNCALLABLE-STANDARD-CLASS, and
>>   (3) The union of (1) and (2).

The most important issues of my previous message were:
 (A) Which of the three classes defined in my previous message 
     does STANDARD-OBJECT refer to?, and 
 (B) What is the name of the other class (the one not named
     STANDARD-OBJECT or FUNCALLABLE-STANDARD-OBJECT)?


The sentence (CLtL2, p. 774) "The class named STANDARD-OBJECT is an instance
of the class STANDARD-CLASS and is a superclass of every class that is an
instance of STANDARD-CLASS except itself" appears imply that if both
class (1) and class (3) exist, then class (1) must be called STANDARD-OBJECT
and class (3) must not be an instance of STANDARD-CLASS.  (Class (3) is
a superclass of class (1) by definition, and class (1) cannot also be a 
superclass of class (3)).  It also allows either class (1) or class (3)
to be called STANDARD-OBJECT as long as the other class does not exist.

Has anyone else written programs that have methods that specialize on 
STANDARD-OBJECT which might be called with arguments that are instances of
FUNCALLABLE-STANDARD-OBJECT?


  Richard Harris