Subject: Wegner
From: "William J. Rapaport" <rapaport@buffalo.edu>
Date: Mon, 22 Mar 2010 11:48:48 -0400 (EDT)

A student writes:
"I am reading the article by Wegner, and he wrote that Milner, in 1975,
noticed that concurrent processes cannot be expressed by sequential
algorithms. But I thought that it was proven that sequential processing
can simulate parallel processing. Am I mistaken?"

Reply:
I haven't read Milner's paper, so I don't know what his proof consists
of. However, although a serial algorithm can *simulate* parallelism,
it can't actually *produce* parallel (or concurrent) behavior. In the
sequential simulation of parallel actions A and B, one of them would
have to be done before the other.

(I've asked some of my colleagues who know more about parallel
computation than I do to respond; when they do, I'll let you all know.)