Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!rutgers!njin!princeton!phoenix!eliot
From: eliot@phoenix.Princeton.EDU (Eliot Handelman)
Newsgroups: comp.lang.lisp
Subject: functionp, fboundp (was: Re: Survey question)
Keywords: reader,type
Message-ID: <4593@phoenix.Princeton.EDU>
Date: 26 Nov 88 20:28:08 GMT
References: <26250@bu-cs.BU.EDU> <3320@mit-amt>
Reply-To: eliot@phoenix.Princeton.EDU (Eliot Handelman)
Organization: Princeton University, NJ
Lines: 25

In article <3320@mit-amt> sokolov@media-lab.media.mit.edu (Michael Sokolov) writes:
>>(setq foo (read))
>>(lambda (x) x)

>	I think the question was what type of object is foo bound to?
>I think it is basically just a list. (fboundp 'foo) returns nil.

Since READ doesn't call EVAL it is just returning a cons. If you need
to have the arguments evaled then you have to write (eval (read)).
That would cause an error in the above example, since LAMBDA is not
a function. 

>however, (funcall foo 1) does return 1.  Accordingly, (functionp foo)
>returns t. I discovered that (functionp 'bar) also returns t, in spite
>of the fact that bar was previously unbound: anyone know why?

FUNCTIONP only says whether its argument is suitable to passed to FUNCALL,
which is true of all symbols. If you want to find out if a symbol has a 
global function definition attached to it you have to use FBOUNDP.

There was some discussion a while back -- though I think it may have been
in the KCl list -- as to what the point of FUNCTIONP was. I think it was
generally agreed that it was pretty pointless. I don't recall whether
X3J13 was planning on rethinking it or not -- perhaps someone from there
might fill us in on the latest update.