Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!tut.cis.ohio-state.edu!pt.cs.cmu.edu!sei!firth
From: firth@sei.cmu.edu (Robert Firth)
Newsgroups: comp.lang.misc
Subject: Re: What is Lazy Evaluation?
Keywords: expression evaluation, lazy evaluation
Message-ID: <6081@bd.sei.cmu.edu>
Date: 13 Feb 90 19:09:13 GMT
References: <3109@caesar.cs.montana.edu>
Reply-To: firth@sei.cmu.edu (Robert Firth)
Distribution: usa
Organization: Software Engineering Institute, Pittsburgh, PA
Lines: 44

In article <3109@caesar.cs.montana.edu> icsu8209@caesar.cs.montana.edu (Glassy) writes:
>What does 'lazy evaluation' mean?  Is it the same thing as 'short-circuit evaluation?

I'll have a shot at an answer.

(a) 'short-circuit' versus 'full' evaluation is usually considered
    a compile time issue, ie the compiler generates one of two
    code sequences.

Example:

	if x>=0 and then sqrt(x)<=y ...

this is short circuit evaluation, something like

	test x
	branch negative L
	call sqrt(x)
	compare y
	branch greater-than L
	...
    L:

so if the first test is enough to decide the condition, the second is
not performed.  By contrast, full evaluation always evaluates everything.

(b) 'lazy' versus 'greedy' evaluation is usually considered an issue
    in implementing language interpreters, eg for lambda calculus.
    Greedy evaluation evaluates an expression as soon as it can; lazy
    evaluation as late as it can.

Example:

	munge(1/x)

greedy evaluation will evaluate 1/x and then call munge().  lazy
evaluation will call munge(), and evaluate 1/x only when the body
of munge referenced its formal parameter.

As the example shows, lazy evaluation "works" in cases where greedy
evaluation does not - in this case, if x=0 but munge doesn't reference
its parameter.

Hope that helps.