Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!hplabs!otter.hpl.hp.com!otter!sfk
From: sfk@otter.hpl.hp.com (Steve Knight)
Newsgroups: comp.lang.misc
Subject: Re: Re^4: Closures (was Re: class-sic.)
Message-ID: <2400030@otter.hpl.hp.com>
Date: 25 Jan 91 09:46:31 GMT
References: <446@data.UUCP>
Organization: Hewlett-Packard Laboratories, Bristol, UK.
Lines: 27

Ken writes:
> I must claim ignorance of Pop, but in common usage [...]
> a closure is called such because it captures free variables by
> "closing over" the environment at time of creation.  This differs from
> currying [see (1), (3)] which seems to be what you mean by "closure".
  ^^^^^^^^
Partial application, not currying, of course.  I presume this is a slip of
the fingers.

Yes, partial application can be used to implement a full-closure mechanism.
It is fairly easy to transform (nested) procedures so that they have no
free-variables EXCEPT for global-variables -- this is akin to lambda-lifting
but works equally well in imperative languages.  This transformation assumes
partial application works, of course.

Having transformed a program into this form, all closure-forming
operations are trivially tackled by partial application, since formal
parameter lists are the only environments that need to be closed over.  This
resolves the conflict of terminology.

> Again, my viewpoint comes from literature in lambda calculus and
> denotational semantics.  Are there references in the general
> literature which support your definition?

As you can see, the same ones :-)

Steve