----- Original Message -----From: Alan KaySent: Thursday, December 04, 2003 4:52 PMSubject: Re: Morph>>DeleteLots of appearance things were nicer before the Smalltalk-80 release efforts. I wasn't there at the time, but I think they decided (probably rightly) to use the existing character set in the outside world. And, today, it would be nice to do something like your suggestion (maybe a different one) and many other cosmetic possibilities, but several folks have pointed out that even a slight deviation from unadorned text is not supported by all email clients (sigh).Cheers,Alan-----At 4:30 PM -0800 12/4/03, Andrew P. Black wrote:Aaron J Reichow wrote:
>
> > > 2. The format for specifying a message is Class>>#message ; that is your
> > > message should have read Morph>>#delete.
> >
> > Really? I use #delete to talk about the message, but Morph>>delete to
> > talk about a particular method.
>
> The convention is to use the #. It's kind of neat actually- do a
> print-it on "Morph>>#delete" in a workspace- the CompiledMethod for that
> method is returned. #delete can be passed (a symbol), and delete cannot.
Yes, "Morph>>#delete" is a valid Smalltalk expression and"Morph>>delete" is not.I've always found the # symbol to be a blot on the otherwise elegant syntax of Smalltalk. Its a bit surprising to me that Smalltalk didn't adopt a different convention, such as typing all symbols in bold face. delete would be an identifier reference, whereas delete would be a symbol, andmyMorph delete would send the message delete to the object bound to myMorph.However, I'm about 25 years too late in proposing this, so I don't expect that it will be adopted real soon now.Andrew --