Dear Krzysztof:

Thank you for your chapter. I am not through yet but here is some
feedback. Overall I like it a lot.

I like your positioning of traversal strategies as an aspect composition
mechanism. page 11, lines 6-7
They can compose more than structure and behavior.
For example, with RIDL they compose structure and the remote invocation aspect.
==

The connection between AOP and AP can be described more
explicitly than you do. AP is a special case of AOP as described in:

http://www.ccs.neu.edu/home/lieber/connection-to-aop.html

=======
With SOP and CP you mention the original publications.
For AP the original publication is:

@INPROCEEDINGS{karl:comp-enh,
AUTHOR = "Karl J. Lieberherr",
BOOKTITLE = "Information Processing '92, 12th World Computer Congress",
TITLE = "Component Enhancement: An Adaptive Reusability Mechanism for Groups of
Collaborating Classes",
YEAR = "1992",
ADDRESS = "Madrid, Spain",
PAGES = "179-185",
EDITOR = "J. van Leeuwen",
PUBLISHER = "Elsevier" }

I have updated:
http://www.ccs.neu.edu/research/demeter/papers/publications.html

accordingly. See top two entries.

=======
In the LoD definition on page 6, line 20 you add: local variables.

We did not have this in the original definition for the following reason:

object.part1().part2().part3()

could be made legal by:

var lv1 = object.part1();
var lv2 = lv1.part2();
var lv3 = lv2.part3();

This would be against the LoD.

From where did you get the above definition of the LoD?

Probably more to come.

-- Karl