From: "Brent Pinkney" <brent.pinkney at aircom.co.za>
> On Tue, 04 Nov 2003 07:27:31 -0800, Michael Rueger
> <michael at squeakland.org> wrote:
> > Brent Pinkney wrote:
> >> You are arguing that we should change the name of MessageNotUnderstood
> >> to MessageNotUnderstoodException or ExceptionMessageNotUnderstood.
> >> Surely by extension we chould change Integer to ClassInteger ans String
> >> to ClassString ?
> >
> > No, because an the integer class is still about integers, whereas the
> > [ x /y ] on: DivideByZeroException do: [ self inform: 'oops' ]
> >
> > it speaks (small-talk) to you ;-)
>
> That is of course your prerogative, but you forget that the precedent of
> voiding Hungarian Notation has already been chosen, MessageNotUnderstood
> and DivideByZero being the cannonical examples.

May I add that I agree with you that this convention exists (and I agree
with the convention), but only inasfar as the convention is "system classes
don't need prefixes or suffixes". This doesn't mean they can't reasonably
have them, just that they don't _need_ them.

> So the other *Exception classes in the image are incorrect.

I agree there too; it is important to be consistent. IBM Smalltalk has
avoided the whole problem from the start by prefixing e.g. the system
exception classes with Ex, as in ExError.

> Whilst no one class is sufficent cause to get overly upset about, enough
> poorly named classes will create sufficient entropy to destroy a large
> portion of the aethetic aspect of Smalltalk.

I agree with you there as well :-).

Cheers,

Peter van Rooijen
Amsterdam