On Sunday, August 25, 2002, at 08:07 PM, Richard A. O'Keefe wrote:

> "Andrew C. Greenberg" <werdna at mucow.com> wrote:
> 	I don't know about this Richard.  #removeAll: implies certain
> 	side-effects with respect to the receiver that render its meaningful
> 	definition undefined when the parameter is modified during its
> 	operation.
>
> I wish people wouldn't use the term "SIDE-effect" for what is the
> MAIN effect of an operation.  It's an important distinction.  Just as
> when a doctor prescribes you a drug, the SIDE effects of that drug are
> almost certainly not intended by her, so when an operation has effects
> other than its main defined purpose, those side effects may well not be
> intended by the user.

"Side-effect," as used here, is a long-standing term of art used in 
imperative and quasi-functional languages to distinguish between the 
returned answer of a message send and the impact of the message send on 
the state of the object.  I am not inclined to adopt new jargon to 
simply for the purpose of sophistic advocacy.