Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: Notesfiles $Revision: 1.6.2.17 $; site uiucdcs.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!inuxc!pur-ee!uiucdcs!kamin
From: kamin@uiucdcs.UUCP
Newsgroups: net.lang
Subject: Re: jnw's question about omitting the 'b
Message-ID: <26400019@uiucdcs.UUCP>
Date: Fri, 14-Dec-84 11:15:00 EST
Article-I.D.: uiucdcs.26400019
Posted: Fri Dec 14 11:15:00 1984
Date-Received: Sun, 16-Dec-84 04:54:50 EST
References: <3748@ucbvax.UUCP>
Lines: 25
Nf-ID: #R:ucbvax:-374800:uiucdcs:26400019:000:1343
Nf-From: uiucdcs!kamin    Dec 14 10:15:00 1984


I am also curious about the omission of "bu", and (forgive me) couldn't
understand Baden's "response."  It does not appear that "bu" was omitted
in favor of a more general currying function - at least I can't find any
in the manual.  (And obviously such a function can't be user-defined, since
it is a functional.)  I'd just like to re-submit this question.

By the way, I would like some clarification on the comment made by jnw,
and endorsed by Baden, on a "more general" currying function. The idea is that
a function f may have a pair of arguments , and you want to obtain,
for fixed x0, a function F such that F:y = f:.  This might be
generalized to f having more than two arguments, but this is quite a
different thing: for fixed x1,...,xn, find F such that F: =
f:.  This is not a "generalization" of bu in the
sense that bu is a special case, because in the two argument case, we
would have F: = f:, and  is not the same as y.  A further
generalization is to allow some subset of the arguments to be fixed, but
this would be a notational nightmare (this kind of thing can be done
nicely only by allowing user-defined functionals).  So I would like to
know what Wilson and Baden have in mind.

				Sam Kamin   (uiucdcs!kamin)
				C.S. Dept.
				U. of Illinois at Urbana-Champaign