Hi Mira:

Luis and I talked about the connection between APPC's and 
Personalities.

The composition discussion is already dated because of your
interface proposal of yesterday.

There is also a note about two ways of expressing the same thing below.
What kind rules do we use reagrding which strategies must result
in unique paths?

Do you agree with this discussion?

Maybe I missed your answer while I was in Germany.

-- Karl

=============================
Hi Mira:

we had a very interesting class yesterday about personalities
and APPCs and Luis and I have worked out some more connections.
Luis will have a better explanation but here is a summary:

Personality			APPC

one class-valued		multi class-valued
variable collaboration		variable collaboration

traversal strategy is simple:	any traversal strategy
a bunch of subclass edges 

micro-architecture 		micro-architecture 
implementation refers		implementation refers 
to methods of the same 		to participants "far away"
class				as specified by a strategy

are a special kind of APPC	
where the collaboration
is within classes and 
subclasses

can be composed			can be composed
                                example:
                                (A -> B -> C -> D -> E}
				{A.i1, A.i2} <- C, {B,C} <- E

corresponds to a role		corresponds to a group of roles


==================================
Luis was noting the following: What is the difference between

Assume:
s = from A to z:Z

A{ void f() 
  during s
    Z{ some code using this }
}
and
A{ void f(){ some code using z} 

I guess the answer is:
If there is a unique path from A to Z (and no
collection class involved) then there is no difference.

We should probably mention this in our paper.


-- Karl