[PRL] AOP questions

[Doug Orleans]

Doug Orleans writes:
 > Matthias Felleisen writes:
 >  > It's fair to say that John should have written AspectJ (i.e., excluded 
 >  > HyperJ and friends).

Also, Hyper/J has dynamic join points, sort of: the "bracket"
integration relationship can insert conditional code that's only
executed when you're being called from some particular other
operation.  The main difference between Hyper/J and AspectJ is that the
former is explicitly about moving code around, while the latter
pretends to be about dynamic execution but is implemented by moving
code around.  (Not that this is a bad thing, it's the standard idea of
implementing an interpreter as a compiler.)

--Doug