Hi Mira:

here are some thoughts on APPCs

APPC V
  Input-Class-Interface
    strategy types
      also introduce participant names
    class-valued-variables between < and >
    method-valued-variables
  Input-Event-Interface
    listens to the following events
  Output-Event-Interface
    produces the following events
{
	instance variables
	methods
	// event based methods, may use wildcard symbol in strategies
	before <ClassName>(Event)
	after <ClassName>(Event)
	vetoFrom<ClassName1>To<ClassName2>(Event)
	veto strategy s (Event)
	before from<className1>To<ClassName2)(Event)
	after from<className1>To<ClassName2)(Event)
	before strategy s (Event)
	after strategy s (Event)
	methodName(Event)
	// several redefinitions:
	redefines <C> { // vertex based redefinition
	  // additional instance variables
	  D d; // not needed for attachment to objects
	       // only for attachment to calls
	  // add new methods for C-objects
	  f(A) {};
	    // "derived" edges (methods)
	  before x(A) {}
	  after x(A) {}
	  around x(A) {}
        }
}


New parts and functions may only be used in redefinition.

We can attach APPCs to:

method calls to temporarily modify the behavior of a group of classes

a class graph to permanently modify a group of classes

anAcc.deposit(a) ::+ LOGGING_PATTERN(...)

will add the log.deposit(a) at the end of Account::deposit(a)
and add a new method deposit to class Logging. 
(It might be strange that the log contains deposit logs although
the deposit method is gone)

classgraph ::+ LOGGING_PATTERN(...)

will update both deposit and withdraw.