>  You know, if Squeak is really slow I wonder if it's not a compiler
>problem.  RICS chips really really like smart compilers :-)

I suspect this is true, by probably a factor of two.  We have honed the Squeak VM in a loop that includes looking carefully at the generated machine code each time we make a major change.

However I found that Squeak actually ran quite usably on the old Cassiopeia and HP320LX (who needs to recompile the system on an airplane?).  The new machines are three times faster than those, and the Squeak VM is 30% faster than it was when I ran those tests.  Even on those early machines, I was more concerned about screen real estate, readability, and pen interface details than about speed.

Finally, I expect Jitter to give us somewhere between a factor of two and three before long, which should be equivalent to bumping the new PDAs up to 200-300MHz ;-).  Of course Jitter will only help Squeak speed -- not BitBlt -- and that takes us back to the original topic of compilers.

	- Dan