Path: utzoo!utgpu!watserv1!watmath!iuvax!mailrus!hellgate.utah.edu!defun.utah.edu!sandra
From: sandra%defun.utah.edu@cs.utah.edu (Sandra J Loosemore)
Newsgroups: comp.lang.lisp
Subject: Re: FUNCALL question
Message-ID: <1990Feb1.212652.10062@hellgate.utah.edu>
Date: 2 Feb 90 04:26:52 GMT
References: <3277@accuvax.nwu.edu> <1990Jan28.175437.19293@hellgate.utah.edu> <1666@skye.ed.ac.uk> <387@forsight.Jpl.Nasa.Gov>
Organization: PASS Research Group
Lines: 41

In article <387@forsight.Jpl.Nasa.Gov> gat@robotics.Jpl.Nasa.Gov (Erann Gat) writes:
>Nevertheless, I stand by the 
>assertion that identifiers in Common Lisp, be they local or global,
>have two bindings, a value binding and a function binding,

Agreed, but they actually have more than two bindings.  They can also
have type bindings, documentation-type bindings, declaration bindings,
and so on. 

>and that
>this is the reason for the existence of an explicit FUNCALL function,
>that is, to enable one to call a function stored in an identifier's
>value slot.

No, the reason for the existence of the FUNCALL function is that in
Common Lisp syntax, the CAR of a list follows different evaluation
rules than subforms in the CDR.  What FUNCALL buys you is the ability
to obtain the function to be called by evaluating something with the
CDR evaluation rules instead of the CAR evaluation rules.  Although
it's hard to imagine why one would want to do such a thing, you could
certainly have a two-namespace Lisp that uses the same evaluation
rules in both CAR and CDR, and no explicit FUNCALL operation.  You
would just have to use #'<function-name> as the CAR of every function
call form instead of just <function-name> so that you refer to the
functional binding of the name instead of its variable binding, as in

    (#'list 'a 'b 'c)

One can also imagine a single-namespace Lisp that has different 
evaluation rules for the CAR than the CDR, that would require an
explicit FUNCALL operator.  The issues of syntax and namespaces are
actually orthogonal.

>I also reaffirm my position that multiple bindings cause a great deal of
>confusion!

I don't think anybody has been seriously taking issue with this.  I'm
not too bothered by multiple namespaces in practice, but I agree that
a single namespace is less complicated.

-Sandra Loosemore (sandra@cs.utah.edu)