Musings about "ma"

David Stes stes at mundivia.es
Fri Oct 16 11:05:13 UTC 1998


Alan Kay wrote:
> 
> Marcel --
> 
> These are just the kinds of directions I was hoping to catalyse. Long ago
> we used to say: find the hardest and most important things you need to do,
> solve them, and then find ways to get everything else out of them. Please
> keep on thinking and writing about these ideas ...
> 
> Cheers,
> 
> Alan
> 
> ------
> 
> At 2:01 PM -0000 10/13/98, Marcel Weiher wrote:

> >Blue plane 1: Software Architecture
> >
> > ...
> >
> >Example 2:  Protocols
> >
> >It would be great to formalize Smalltalk protocols along the lines
> >of Objective-C

Well it's not a very good example that was catalyzed here, because this is
a very controversial extension that is not supported by the majority of
the Objective-C compilers. 

I am personally in doubt whether to add this extension to my Objective-C
preprocessor (see http://cage.rug.ac.be/~stes/compiler.html) and I asked to
Stepstone whether they were going to this, to which a compiler engineer at
Stepstone (see http://www.stepstn.com) replied something like "we call
that tinkering with the language", so they don't seem to happy with it. 

If you change an existing system, whether it's protocols for Objective-C,
or prototypes for Squeak, I think you always have to be careful, that it's
with the consent of the people who built the system.

Otherwise, just be consistent, and develop a completely new system (such
as Self was prototype oriented in my understanding, _based_ on ST but not
claiming to _be_ ST).  Nobody's forbidding anyone to develop a completely
new pure prototype oriented system, but I'm reading this list because I
like Squeak the way it is, as a class-based system.





More information about the Squeak-dev mailing list