[PRL] aspects to eliminate aspects

[Doug Orleans]

Mitchell Wand writes:
 > RF> Can't you just avoid calling proceed in the advice that advises other
 > RF> advice?
 > 
 > No, if the interfering advice was an around advice, then that would
 > skip both the advice and the method call that invoked it.  (Right?)

It might be nice to have a sort of "directed proceed", analogous to a
"directed resend" in OOP that lets you skip some superclasses and go
straight to a particular ancestor.  The problem is, again,
obliviousness-- each aspect doesn't necessarily know what other
aspects are around, so you could only skip all the way to the base
code.  Which is meaningless in a symmetric system like HyperJ or
Socrates where there is no distinction between "base" and "aspect".

--Doug