Hi Crista:

thank you
===============================
Hi Mira:

>From mira@informatik.uni-siegen.de Wed Nov  4 11:22:49 1998
>Subject: Re: qualified visits
>To: lieber@ccs.neu.edu (Karl Lieberherr)
>Date: Wed, 4 Nov 1998 17:22:30 +0100 (MET)
>From: "Mira Mezini" <mira@informatik.uni-siegen.de>
>Cc: mira@informatik.uni-siegen.de, lorenz@ccs.neu.edu
>
>Hi Karl:
>
>I also think that this is useful (please would you keep track of
>all things we have in our agenda. I could do that, but you know
>I am not very organized :-) ). 

I keep track of those things. Maybe we should make a web page:
open issues with APPCs: project and thesis topics.

>
>As far as simple ICG edge qualifiers are concerned, I think 
>they are there: we can define wrappers on ICG edges just like

You are right! Forgot about that.

>those on ICG nodes. In the capacity checking example it happens 
>that those are not needed, but they are useful for the graph example,
>for instance. Soon, I will write this example too.
>
>As for strategies over the ICG as qualifiers, I have the feeling that
>they also might be useful, but we haven't had an example for that yet.
>Need to think about such an example and about how it can be realised.
>It seems interesting.

Maybe we can do most of what we want with: ICG edge qualifiers.
>
>What I also find interesting is to get some (or all) of the examples
>in the Aspect/J tutorial and see how they would be coded with 
>the APPCs. I was going through the tutorial yesterday and was thinking
>that an AJ static aspect is like an APPC with the APPCs being better in
>my eyes (at least when all we plan to do is there). But the examples
>coded in AJ could give us ideas for additional construct we might 
>need. Beside that we would extend our example pool - and maybe have our 
>own tutorial at some time .... Do you have openings in your projects?

A very good idea. I try to find a student in my course.
Both AspectJ and APPCs are about spreading Java code across several existing classes.
But APPCs have some meat: a good composition approach and structure-genericity.
No wonder you like APPCs better :-)
Certainly we can learn something from AspectJ but the AspectJ team can probably
learn more from APPCs.

-- Karl

>
>It's so much to do ... 

We don't have to do it all ourselves. At some point many others will jump on the
APPC bandwagon.

>
>- Mira
>
>> Hi Mira:
>> 
>> thank you for the related work you found. I will get back to that later.
>> 
>> We talked earlier about qualifying wrappers with interface
>> class graph edges in APPCs.
>> 
>> after B with {A1->B} { some code 1 }
>> after B with {A2->B} { some code 2 }
>> 
>> If we got to B from A1, we execute code 1 and if we got 
>> to B from A2, we execute code 2.
>> 
>> Do we still want this or even something more general where
>> the edge is replaced by a strategy over the interface class graph.
>> 
>> I think this is useful and one of our graph examples used it.
>> 
>> -- Karl
>> 
>
>