JArchibald at aol.com affirms
that he uses 'binary message' not for messages that are binary,
but for messages that are spelled without letters, and writes:
	
	What is confusing is to introduce concepts from other languages, and expect 
	them to do as good a job in describing Smalltalk programs as Smalltalk's own 
	concepts.

To be honest, I hadn't thought of "operator" as a programming language concept
at all.  It certainly has nothing to do with precedence, or any particular
programming language.  And if you'd tried teaching Smalltalk to people who
had recently learned about the difference between I mode and A mode in FTP,
you might not find "binary message" so intuitive to them.

Let's face it, "number" is a concept from other languages too,
and if "Boolean" wasn't in Algol before it was in Smalltalk I must have
hallucinated every copy of the Algol 60 report I've ever seen.
Nor is "stream" a concept peculiar to Smalltalk.

What matters is not whether terms or concepts are used in describing
other languages or are peculiar to Smalltalk, but whether they *work*.
Since unary/binary/keyword is not a TRUE description (keyword messages
may be unary and they may be binary), the simple fact that that terminology
WAS used in Smalltalk and IS used by some Smalltalkers does NOT mean that it
is appropriate for all Smalltalkers for all time.

Note that I am not advocating that everyone use the term "operator" for
(historically-so-called) "binary selectors" (not messages!), nor am I
advocating a wholesale revision of Smalltalk terminology.

What I *am* doing is defending my use of "operator" as being at least as
truthful as "binary message" (more truthful, in fact, since #+ is not a
message, it's a selector) and as appropriate in context.

If anyone wants to argue about this any more, please don't.
We have better things to do in this mailing list.