| >| No, I don't think so. I will do some tests on a bunch of third-party
| >| programs, and if they work then that should be it :-)
| >
| >Random testing is not sufficient to prove that this change is innocuous.
| >
|
| Neither is the reverse necessarily true (throwing hands in the air and
| claiming "it might break things!").

On the other hand, when someone is advocating changing the way processes
are laid out in memory, I would at the very least expect an investigation
of what could break and a description of what steps have been taken to
prevent such breakage.  Especially since this may backwards compatibility
with legacy binaries that cannot be simply recompiled.

Eduardo