"Withers, Robert" <rwithers at quallaby.com> is widely believed to have written:

> Tim,  I don't agree.  Don't rely on books so much {raspberry sound effects}
> ;-)
What!?! Sacriledge!
> Consider that in:
> 
> | m n  |
> n := 2.
> m := 3.
> n := m + (m := n).
> 
> m is the receiver of the plus message and not another argument, as in your
> example, so it gets pushed onto the stack before the arguments are
> evaluated, inline.
I don't care what is actually happening now; my point is that it is
_wrong_ to have code work in a way that is so absurdly confusing. I can
see just what the bytecode stream is doing and thus why the answer is
currently what it is but that just ain't a good excuse. Evidence of past
foolishness is not evidence for continued foolishness.

Does anyone really think that it is _proper_ that
| m n  |           
n := 2.            
m := 3.            
n := m + (m := n).

gives a different answer to
| m n  |           
n := 2.            
m := 3.
m := n.            
n := m + m.
??

tim
-- 
Tim Rowledge, tim at sumeru.stanford.edu, http://sumeru.stanford.edu/tim
Don't let the computer bugs bite!