In article <PGEPJIFLPEPOHCKEEEIKAEHBDCAA.james / rubyxml.com>,
 james / rubyxml.com wrote:

> > In article <PGEPJIFLPEPOHCKEEEIKEEGLDCAA.james / rubyxml.com>,
> >  james / rubyxml.com wrote:
> >
> > > Am I being pedantic, or wrong?
> >
> > Maybe just a little of both.  ;-)
> >
> > The problem, as I see it is that every language(that does) tends to
> > implement OO in slightly different ways.  Thus I like to break things
> > down into just a few basic concepts that apply across the spectrum, and
> > deal with implementation quirks when they become an issue.
> >
> > For instance, if I were using Eiffel to convey introductory OO concepts,
> > do I bring Eiffel's powerful inheritance structures into the mix when
> > I'm talking about general theory?  No.  Because, for better or worse,
> > that power isn't very common.
> 
> So, then, the message-passing paradigm is not a general OO concept?
> Or, it is, but not one worth emphasizing, because too few languages
> actually behave that way?

Not at first.  Especially for students with a background in a procedural 
language or languages, there's plenty to trip them up already. 

It took me quite a while until I "got" objects and I remember struggling 
with it(until elightenment in the form of Python hit me like a 16 ton 
anvil), so I try to avoid information overload.