From mira@ccs.neu.edu Fri Jan 23 16:11:41 1998
Subject: Re: Interesting message from M. Felleisen
To: matthias@rice.edu
Date: Fri, 23 Jan 1998 16:11:39 -0500 (EST)
Cc: kanderso@zima.bbn.com, lieber@ccs.neu.edu (Karl Lieberherr)

Hi Matthias:

My name is Mira Mezini and I am currently working with Karl 
Lieberherr at Northeastern.
I just read your paper "Classes and Mixins", which I liked
a lot. Well written.

It seems like we should intense our communicaton, since
our research interests are very close to each other. In my
PhD thesis "Variation-Oriented Programming Beyond Classes
and Inheritance" which I finished in Germany just before moving to
Northeastern, I am pretty much concerned with the same issues you
deal with in your paper. 

There are two main differences in our approaches:

a) In my model (called Rondo) I use the term "adjustments" 
   for what you call "mixins" :-)

b) I don't restrict myself to static combinations of 
   adjustments; Rondo also provides for dynamic combinations 
   to allow the behavior of individual objects to vary
   at run-time.

   
If you want to have a look at Rondo, here are two references:   

There was a paper in ECOOP '97
        Dynamic Object Evolution Without Name Collisions

        you can get an electronic version of the paper in
                http://www.ccs.neu.edu/home/mira/public.html
 
However, I would prefer you read parts of the thesis rather
than the ECOOP paper, for two reasons. The material in the thesis
is more recent and much better elaborated than in the paper. 
More importantly, the emphasis of the paper is on dynamic 
variation, although I notice there that the model is the 
same applicable for static combinations.
The thesis is electronically available in:

        http://www.ccs.neu.edu/home/mira/thesis.html
 
You may want to have a look at chap. 3 and 4, the latter being more 
important. 

It would be also interesting to compare our formal descriptions.
To be honest, I skipped that part of your paper, during the first
reading, but plan to go back to it soon.

There are two other related works I missed in your related works
section :-)

        1) a paper by Prehofer, C., also published in the ECOOP '97
           proceedings, called "Feature-Oriented Programming"
           He also extends Java.

        2) VanLimbergen and Mens
           "Encapsulation and Composition as Orthogonal Operations
           on Mixins: A solution to multiple inheritance problems"
           In Object-Oriented Systems 3(1), 1996


regards,

- Mira