Morphic and eToys

tblanchard at tblanchard at
Wed Feb 12 12:02:05 UTC 2003

I'm still digging.  Last night I worked through part of Alan Kay's map 
application tutorial - the bit where you make continents zoom when you 
mouse over them.

I learned some new things.  Clarified some ideas, and firmed up a 
couple ideas I have.  I also hit a couple walls.

BUG:  Morph top: newTop just sets the x coordinate of position and 
moves the whole morph.  There is no control for adjusting where the top 
is while leaving everything else where its supposed to be.  There are 
more issues with the geometry code I think.  Something that could be 
worked on all by itself.

But thats a nit.

My other observation is that Player is just a delegate.  When a morph 
starts stepping, it thoughtfully forwards the step calls to the player. 
  Fine, but Player itself could just as easily be the thing that gets 
put into the scheduler.  I need to do more digging but it seems to me 

1) Players can be generalized to work on any kind of object, not just 
2) Morphs could be generalized to have a delegate - one kind of which 
might be a player, but might also be another UI component or some other 
kind of model.
3) A little effort generalizing these mechanisms could result in making 
more kinds of objects playable and allowing for more traditional UI 
development without having players if that's what the developer wants.

This is my current thinking.  Still need to do more digging.

Todd Blanchard

More information about the Squeak-dev mailing list