<html><head><meta http-equiv="Content-Type" content="text/html charset=windows-1252"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;">I stand corrected.<div><br><div><div>On Mar 5, 2021, at 10:24 AM, Shannon Spires <<a href="mailto:svs@bearlanding.com">svs@bearlanding.com</a>> wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite">
  
    <meta http-equiv="Content-Type" content="text/html;
      charset=windows-1252">
  
  <div text="#000000" bgcolor="#FFFFFF">
    See comments inline.<br>
    -SS<br>
    <br>
    <div class="moz-cite-prefix">On 3/5/21 9:25 AM, Ron Garret wrote:<br>
    </div>
    <blockquote type="cite" cite="mid:D848EA75-B628-46F7-BA96-0AEEE11FC583@flownet.com">
      <meta http-equiv="Content-Type" content="text/html;
        charset=windows-1252">
      <br>
      <div>
        <div>On Mar 4, 2021, at 10:50 PM, Shannon Spires <<a href="mailto:svs@bearlanding.com" moz-do-not-send="true">svs@bearlanding.com</a>>
          wrote:</div>
        <br class="Apple-interchange-newline">
        <blockquote type="cite">
          <meta http-equiv="Content-Type" content="text/html;
            charset=windows-1252">
          <div text="#000000" bgcolor="#FFFFFF"> > So it makes no
            sense to create an instance of COUNTED-CLASS which inherits
            from RECTANGLE.<br>
            <br>
            Sure it does.</div>
        </blockquote>
        <div><br>
        </div>
        <div>No, it doesn’t :-)</div>
        <br>
        <blockquote type="cite">
          <div text="#000000" bgcolor="#FFFFFF">Any given class
            metaobject has two kinds of relationships with other class
            metaobjects: It has an INSTANCE-OF relationship with its
            metaclass and an ISA (or INHERITANCE) relationship with its
            superclasses.<br>
          </div>
        </blockquote>
        <div><br>
        </div>
        <div>Yes.</div>
        <br>
        <blockquote type="cite">
          <div text="#000000" bgcolor="#FFFFFF"> The equivalent (and
            more conventional) definition of <font face="monospace">counted-rectangle</font>
            would be<br>
            <font face="monospace"><br>
              (defclass counted-rectangle (rectangle)<br>
                ()<br>
                (:metaclass counted-class))</font><br>
            <br>
            and this makes perfect sense.<br>
          </div>
        </blockquote>
        <div><br>
        </div>
        <div>Yes.  And it works:</div>
        <div><br>
        </div>
        <div>
          <blockquote type="cite">
            <div>Welcome to Clozure Common Lisp Version 1.11-r16812M
               (DarwinX8664)!</div>
            <div>? (defclass counted-class (standard-class)</div>
            <div>   ((counter :initform 0)))</div>
            <div>#<STANDARD-CLASS COUNTED-CLASS></div>
            <div>? (defclass counted-rectangle (rectangle)</div>
            <div>  ()</div>
            <div>  (:metaclass counted-class))</div>
            <div>#<COUNTED-CLASS COUNTED-RECTANGLE></div>
            <div>?<br>
            </div>
          </blockquote>
        </div>
      </div>
    </blockquote>
    <br>
    You haven't shown the definition of rectangle above. Let's include
    that:<br>
    <div><br>
      <font face="monospace">Clozure Common Lisp Version 1.12-dev
        (v1.12-dev.5-22-g0430f8db) DarwinX8664<br>
        <br>
        For more information about CCL, please see
        <a class="moz-txt-link-freetext" href="http://ccl.clozure.com/">http://ccl.clozure.com</a>.<br>
        <br>
        CCL is free software.  It is distributed under the terms of the
        Apache<br>
        Licence, Version 2.0.<br>
        ? (defclass counted-class (standard-class)<br>
           ((counter :initform 0)))<br>
        #<STANDARD-CLASS COUNTED-CLASS><br>
        ? (defclass rectangle ()<br>
            ((height :initform 0.0 :initarg :height)<br>
             (width :initform 0.0 :initarg :width)))<br>
        #<STANDARD-CLASS RECTANGLE><br>
        ? (defclass counted-rectangle (rectangle)<br>
          ()<br>
          (:metaclass counted-class))<br>
        > Error: The class #<STANDARD-CLASS RECTANGLE> was
        specified as a<br>
        >        super-class of the class #<COUNTED-CLASS
        COUNTED-RECTANGLE>;<br>
        >        but the meta-classes #<STANDARD-CLASS
        STANDARD-CLASS> and<br>
        >        #<STANDARD-CLASS COUNTED-CLASS> are
        incompatible.<br>
        > While executing: #<a class="moz-txt-link-rfc2396E" href="ccl::STANDARD-KERNEL-METHODCCL::ENSURE-CLASS-INITIALIZED(CCL::SLOTS-CLASS)"><CCL::STANDARD-KERNEL-METHOD
        CCL::ENSURE-CLASS-INITIALIZED (CCL::SLOTS-CLASS)></a>, in process
        Listener(4).<br>
        > Type cmd-. to abort, cmd-\ for a list of available
        restarts.<br>
        > Type :? for other options.<br>
        1 > </font><br>
      <br>
    </div>
    Oops. It doesn't work if rectangle's definition doesn't include <font face="monospace">(:metaclass counted-class).</font><br>
    <br>
    But your point is valid because I didn't test for the case of
    defining <font face="monospace">rectangle</font> <i>after</i>
    defining <font face="monospace">counted-rectangle</font>. That
    works even when <font face="monospace">rectangle</font> does <i>not</i>
    include <font face="monospace">(:metaclass counted-class)</font>.
    The order of class definitions shouldn't matter; I think that's
    probably a bug.<br>
    <br>
    <blockquote type="cite" cite="mid:D848EA75-B628-46F7-BA96-0AEEE11FC583@flownet.com">
      <div>
        <div>But that is not the same thing as creating “an instance of
          COUNTED-CLASS which inherits from RECTANGLE”:</div>
        <div><br>
        </div>
        <div>
          <blockquote type="cite">
            <div>? (class-of (find-class 'counted-rectangle))</div>
            <div>#<STANDARD-CLASS COUNTED-CLASS></div>
          </blockquote>
          <blockquote type="cite">
            <div>? (class-direct-superclasses *)</div>
            <div>(#<STANDARD-CLASS STANDARD-CLASS>)</div>
            <div>? </div>
          </blockquote>
          <br>
        </div>
        <div>In this case, COUNTED-CLASS inherits only from
          STANDARD-CLASS, not from RECTANGLE.  And that does indeed make
          sense.</div>
      </div>
    </blockquote>
    <br>
    This is where it's important to use precise terminology.<br>
    <br>
    COUNTED-RECTANGLE is literally "The class metaobject named
    COUNTED-RECTANGLE which happens to be an instance of COUNTED-CLASS,
    and which also happens to inherit from RECTANGLE."  This makes
    perfect sense when one understands that COUNTED-RECTANGLE is a class
    metaobject, because only class metaobjects can "inherit" from
    anything. (Ordinary CLOS objects that are <i>not</i> class
    metaobjects cannot "inherit" from anything. They are merely
    instances of a single class, period. Although we might loosely say
    things like "the LION named Sam inherits from CARNIVORE and
    SAVANNAH-RESIDENT" that phrasing is not technically accurate; only
    the class LION can inherit from CARNIVORE and SAVANNAH-RESIDENT. As
    an instance of LION, Sam has no "inheritance" relationship with
    anything (except perhaps his ancestor lions, but that's not the kind
    of "inheritance" we're talking about here.))<br>
    <br>
    The <font face="monospace">:direct-superclasses</font> argument in
    the AMOP example specifies the inheritance relationship for class
    metaobjects. It expresses the same thing as <font face="monospace">(rectangle)</font>
    below:<br>
    <br>
    <font face="monospace">(defclass counted-rectangle (rectangle)<br>
        ()<br>
        (:metaclass counted-class))</font><br>
    <br>
    Which again makes perfect sense.<br>
    <br>
    <blockquote type="cite" cite="mid:D848EA75-B628-46F7-BA96-0AEEE11FC583@flownet.com">
      <div>
        <blockquote type="cite">
          <div text="#000000" bgcolor="#FFFFFF"> What doesn't work in
            the example is that <font face="monospace">rectangle</font>
            and <font face="monospace">counted-rectangle</font> have
            different metaclasses.</div>
        </blockquote>
        <div><br>
        </div>
        <div>No, that’s not the problem.  In the example as you’ve given
          it (not as the MOP book gives it), COUNTED-RECTANGLE has a
          metaclass of COUNTED-CLASS, and RECTANGLE (assuming you define
          it as the MOP book does) has a meta-class of STANDARD-CLASS.
           And all that works just fine.</div>
      </div>
    </blockquote>
    <br>
    It only works if you define RECTANGLE <i>after</i>
    COUNTED-RECTANGLE. Which I think is probably a bug.<br>
    <br>
    <blockquote type="cite" cite="mid:D848EA75-B628-46F7-BA96-0AEEE11FC583@flownet.com">
      <div>
        <div><br>
        </div>
        <div>The problem with the MOP example:</div>
        <div><br>
        </div>
        <div>
          <div text="#000000" bgcolor="#FFFFFF">
            <blockquote type="cite">
              <div class="" style="word-wrap: break-word;
                -webkit-nbsp-mode: space;"><span class="" style="font-family: Monaco;">? (setf (find-class
                  'counted-rectangle)</span><br class="" style="font-family: Monaco;">
                <span class="" style="font-family: Monaco;">       (make-instance
                  'counted-class</span><br class="" style="font-family:
                  Monaco;">
                <span class="" style="font-family: Monaco;">         :name
                  'counted-rectangle</span><br class="" style="font-family: Monaco;">
                <span class="" style="font-family: Monaco;">         :direct-superclasses
                  (list (find-class 'rectangle))</span><br class="" style="font-family: Monaco;">
                <span class="" style="font-family: Monaco;">         :direct-slots
                  ()))</span></div>
            </blockquote>
            <br>
          </div>
          <div text="#000000" bgcolor="#FFFFFF">Is that it is trying to
            make the metaclass of COUNTED-CLASS be RECTANGLE, and that
            doesn’t work because RECTANGLE is not a metaclass,</div>
        </div>
      </div>
    </blockquote>
    <br>
    Nope. Again, <b><font face="monospace">:direct-superclasses</font></b><b>
      is about the superclasses of the class metaobject; it is not about
      the metaclass of the class metaobject.</b><b><br>
    </b><br>
    <span class="" style="font-family: Monaco;"></span><br>
    <blockquote type="cite" cite="mid:D848EA75-B628-46F7-BA96-0AEEE11FC583@flownet.com">
      <div>
        <blockquote type="cite">
          <div text="#000000" bgcolor="#FFFFFF"> Any time class A
            inherits from class B, they both should be instances of the
            same metaclass.</div>
        </blockquote>
        <div><br>
        </div>
        <div>That is correct, but you’ve misapplied the rule.  The
          problem is not that "rectangle and counted-rectangle have
          different metaclasses”.  The problem in this case is that
          RECTANGLE and *COUNTED-CLASS* (not counted-rectangle) have
          different metaclasses.  Having rectangle and counted-rectangle
          have different metaclasses works perfectly well, as the
          example above demonstrates.</div>
        <div><br>
        </div>
        <div>Another way to illustrate what the problem actually is
          here:</div>
        <div><br>
        </div>
        <div>
          <div>? (typep (make-instance 'standard-class) 'class)</div>
          <div>T</div>
          <div>
            <div>? (typep (make-instance 'counted-class) 'class)</div>
            <div>T</div>
          </div>
          <div>? (typep (make-instance 'rectangle) 'class)</div>
          <div>NIL</div>
          <div><br>
          </div>
          <div>That is the reason that it makes no sense for RECTANGLE
            to be a DIRECT-SUPERCLASS of an instance of COUNTED-CLASS.
             An instance of COUNTED-CLASS is a class, but an instance of
            RECTANGLE is not.  So an instance of the resulting chimera
            would have to be both a class and not a class.</div>
          <div><br>
          </div>
          <div>But yeah, it gets pretty confusing.</div>
          <div><br>
          </div>
        </div>
        <div>rg</div>
        <div><br>
        </div>
      </div>
    </blockquote>
    <br>
  </div>

</blockquote></div><br></div></body></html>