From: Nicolas Neuss
Subject: Question concerning AMOP
Date: Thu, 08 Jan 2004 10:28:32 +0000
Message-ID: <87brpeybsf.fsf@ortler.iwr.uni-heidelberg.de>

Hello.

I have a question to AMOP, Page 204 (Initialization of Slot Definition
Metaobjects, see <http://www.lisp.org/mop/dictionary.html#Initialization>)

  The :initfunction argument is a function of zero arguments (?) which,
  when called, evaluates the :initform in the appropriate lexical
  environment. The :initfunction argument defaults to false. An error is
  signaled if the :initfunction argument is supplied, but the :initform
  argument is not supplied.

Should this read "of one argument"?  Or should initform and initfunction be
independent?

Nicolas.

• Re: Question concerning AMOP Frode Vatvedt Fjeld
  • Re: Question concerning AMOP Pascal Costanza

From: Frode Vatvedt Fjeld
Subject: Re: Question concerning AMOP
Date: Thu, 08 Jan 2004 11:25:27 +0000
Message-ID: <2had4yiswo.fsf@vserver.cs.uit.no>

Nicolas Neuss <·······@iwr.uni-heidelberg.de> writes:

>   The :initfunction argument is a function of zero arguments (?) which,
>   when called, evaluates the :initform in the appropriate lexical
>   environment. The :initfunction argument defaults to false. An error is
>   signaled if the :initfunction argument is supplied, but the :initform
>   argument is not supplied.
>
> Should this read "of one argument"?  Or should initform and
> initfunction be independent?

No, zero arguments is correct, because the initfunction is a closure
that closes over all the information it needs. It'll be something like
this:

  :initform 123
  :intifunction (lambda () 123)

Or

(let ((lexically-other 'blargh))
  ...

    :initform (compute 'foo 42 (1+ *something*) (1- lexically-other))
    :initfunction (lambda ()
                    (compute 'foo 42 (1+ *something*)
                             (1- lexically-other)))
  ...)

So they are not independent, the form is just the symbolic version of
the intifunction. I'm not actually sure it's really required/necessary
to keep the (symbolic) initform around.

-- 
Frode Vatvedt Fjeld

From: Pascal Costanza
Subject: Re: Question concerning AMOP
Date: Thu, 08 Jan 2004 12:16:09 +0000
Message-ID: <btjhmb$b1s$1@newsreader2.netcologne.de>

Frode Vatvedt Fjeld wrote:

> So they are not independent, the form is just the symbolic version of
> the intifunction. I'm not actually sure it's really required/necessary
> to keep the (symbolic) initform around.

I don't think so because in general, the initform doesn't provide enough 
information to recreate the correct initialization value. I guess it's 
main purpose is to provide some form of documentation.


Pascal

-- 
Tyler: "How's that working out for you?"
Jack: "Great."
Tyler: "Keep it up, then."