better or worse (was: YASoB)
alan.kay at squeakland.org
Thu May 18 19:11:54 UTC 2006
At 11:44 AM 5/18/2006, Jecel Assumpcao Jr wrote:
> > Jecel ,
> > >
> > > My comment on this is that Smalltalk-80 was indeed wonderful by relative
> > > standards, but it shouldn't become a religion that keeps us from
> > > inventing something better. Though this isn't nearly as sad as people
> > > who keep insisting on creating things that are worse while the public
> > > assumes it is automatically better than something "old" like Smalltalk
> > > (what C. S. Lewis called "chronological snobbery").
> > >
> > I have a feeling that to many Smalltakers, in general, there have been
> > no advances in software engineering and computer language design
> > since Smalltalk was invented.
>Lispers feel the same way. We have to be careful not to make the
>opposite mistake ("only the old stuff was good") which is just another
>form of chronological snobbery.
> > When was Smalltalk really invented?
> > Was it in 1972 or 1976 or 1980?
>In 1972 we got the seed (objects and messages) and in 1976 the current
>form (vm, classes as objects, inheritance).
> > Did Squeak Central insist on creating things that are worse than
> > Smalltalk-72 and the crowd assumed that it is automatically better
> > than something "old" like Smalltalk-72?
>"Better" and "worse" don't always apply - I can't compare Smalltalk with
>Prolog, for example. But I can compare Smalltalk with Java or Ruby even
>though it is still complicated even in these cases.
>As far as I know the goal of Squeak Central was to create the eToys
>system and they needed a universal platform to build it on. They had no
>wish to spend their time on such a platform so given the hype around
>Java back then they considered using that. The shocking part was that
>even though it was much newer and borrowed some aspects from Smalltalk
>it was in general "worse". In part this was on purpose ("oh.... the
>average programmer can't understand or handle that...") and in part it
>was ignorance (Guy Steele was brought in late in the project to help
>with the documentation and ended up fixing some minor technical problems
>but had to leave the major ones alone) but it was felt at the time that
>doing eToys in Java would be a larger effort than cleaning up the old
>Apple Smalltalk and using that. If my impressions of Squeak and Java
>history are wrong I hope someone will correct me.
>Note that John Maloney came to Squeak Central from the Self project.
>Some people (including me) feel that Self is better than Smalltalk-80
>but John said in an interview that going back to the older system made
>him more comfortable (I understood him to mean specially the "whole
>system" view that the browser gives you compared to the more piecemeal
>views that you get with Self's outliners). And there are tradeoffs: when
>Logo was created from Lisp it dropped closures (making it worse) because
>they got in the way of the language's goals.
>What I am saying is: imagine that someone is creating a new language
>today. And then you look at what they are doing and give them a good
>demo of APL. They might say "I considered that, but in this part of my
>language...." or (which is far more common) they might say "Wow! I never
>imagined there could be something like that!"
>Certainly the latter never was the case for Squeak Central.
> > For that matter, did the commercial Smalltalk vendors insist
> > likewise?
>Vendors are by their very nature conservative (their customers value
>stability). David Simmons (SmalltalkAgents, S#) seems to me to be the
>only one from that part of the world who is willing to explore new
More information about the Squeak-dev