"Andrew C. Greenberg" <werdna at mucow.com> wrote:
	Stephane's initial proposal was certainly written in at least as 
	provocative a tone as was Richard's reply, but neither message was 
	wholly out-of-line.

If anyone perceived anything "provocative" in my message, I am
(a) sorry, and
(b) AMAZED.
It was a straightforward reductio ad absurdum reply to a proposal to
break a lot of people's code.  There was nothing intended as ad hominem
or impolite in it, and on reviewing it, I don't see anything ad hominem
or impolite.

	(On the other hand, Dijkstra fan that I am, I also missed losing the 
	form of multiple assignment permitted in earlier images, using "{a. b. 
	c} := {exp1. exp2. exp3}".)
	
Me too.  For the life of me I can't see why the Java designers, for
example, elevated "a method must return at most one result" from an
inconvenience to a pillar of the OO faith.  I don't particularly care
about the syntax, but multiple results (as in Lisp and R5RS Scheme)
and concurrent assignment are appropriate in some designs.  I note that
Eiffel, which has some very strong opinions behind it, has recently added
"TUPLE"s as a way of expressing ad hoc aggregations that don't really
warrant a class of their own.