[Edu-sig] Entering Squeakland

Scott David Daniels Scott.Daniels at Acm.Org
Sat Mar 11 21:45:58 CET 2006


kirby urner wrote:
>> I think you owe it to yourself, and perhaps to us, to understand  and
>> express why Squeak does not represent the perfect environment for pursuing
>> the kind of educational ideas that you tend to express. If it in fact does
>> not....
>> My own concerns start exactly there - with the word "environment".
>>
>> My understanding is that many conclude that too much "environment" is what
>> doomed Smalltalk to a  minor role in today's software world.
> 
> Yes, there's lots to say about this (most of it not by me).  In
> SmallTalk what you save is "the image" -- which is the world as you've
> massaged it and messaged it up until now.  You pass whole worlds
> around, more than fragments thereof.  Context becomes everything.

This is both the strength and weakness of Smalltalk:
Then environment is a single, learnable model.  You can look into the
guts and see how it works.  Yo can change the way the system works and
see the effects of the change.  This is gold for the tinkerers among us. 
  What techie-kid has not wanted to take things apart and see how they
work with a bit of a twist?

It becomes hard to build separate things that work well with anybody's
hallucination of what the core is.  This is a problem with Lisp and Ruby
systems as well.

The other problem is the "one world model" which Smalltalk shares with
Lisp, APL, and Prolog (and many others).

> My guess is SmallTalk wizards long ago added the atomism people were
> missing, but by that time, the other language designers had already
> absorbed the OO paradigm, and were reimplementing it in C (to give us
> C++, Java, C#) and in Python.

But many shared Smalltalk things involve mucking with the core objects:
"This is my nifty whizz-bang: just add these methods to 'object' and
then you'll find that ...."


-- Scott David Daniels
Scott.Daniels at Acm.Org



More information about the Edu-sig mailing list