> Note that SOAR never actually ran any applications, just tiny tests. By
> the end of the project it had become obvious that dynamic translation
> (Jitter) of bytecodes was more practical than native compilation. This
> lesson was applied to VisualWorks, Mushroom and Self.

Could you elaborate on this?  The only reason why bytecodes would be more
practical seems to be that the code is recompilable if you don't have the
source any more.

For a dynamic compiler to be effective, it would have to rebuild the
abstract syntax tree anyhow -- why not from the source?