ECOOP 2005 Technical Papers Notification
From: "Andrew P. Black" <ecoop@cs.pdx.edu>
To: christian.heller@tu-ilmenau.de, detlef.streitferdt@tu-ilmenau.de, ilka.philippow@tu-ilmenau.de
CC: ecoop2005-papers-webadmin@borbala.com
Date: 22.02.2005 00:39

Dear Christian, Detlef and Ilka,

It is with regret that I write to inform you that your paper
"A new Pattern Systematics"
has not been accepted for presentation at ECOOP 2005.

The selection process was highly competitive; out of 172 submissions,
the committee accepted only 24 papers, a little less than 14 per
cent. Over 500 reviews were written by members of the committee and
by outside reviewers. I believe that these reviews will provide you
with useful information to improve your paper.

Your reviews are included below.

Although I am sure that you will be disappointed that this paper was
not accepted, I believe that ECOOP 2005 will be a strong conference,
and hope that you will be able to attend, hear the selected papers,
and enjoy camaraderie of the conference and the sights of the city of
Glasgow.

    Andrew P. Black (Prof)
    ECOOP 2005 Programme Chair

*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=

First reviewer's review:

     >>> Summary of the submission <<<

The paper tries to propose a new organization of design patterns

     >>> Evaluation <<<

I will try to help you the paper:

- Patterns are about naming and tradeoffs.
- I do not understand why you want to eliminate the repetition of patterns
if this is needed this is needed.

- what is exactly the goal of section 2.3? What is the point to repeat these
patterns? And summarizing patterns is reducing the essence of patterns (the
tradeoffs) to useless information.

- Problems section arrives faaaaar too late page 9!

- I suggest you to read the ObjVlisp paper of Pierre Cointe OOPSLA 87
to understand the metaclass architecture

- I do not understand the section 3.1 and what is the point of the section on
the JVM

*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*

Second reviewer's review:

     >>> Summary of the submission <<<

A way to categorize or think about software patterns, based on the way human
cognition works.

The paper consists of an analysis of what software patterns are and a
recapitulation of a number of well-known ones.

Finally, the systematics is introduced.

     >>> Evaluation <<<

This work may one day be ready for ECOOP, but not yet. The authors are in the
midst of trying to make sense of the concept of patterns as described by
writers in the area while looking at a fairly large slice of a certain branch
of the realm. The theory as explained so far is immature - the authors need to
take the work deeper and also show why their new systematics makes a
difference. What is the insight and what are its benefits. What is the
significant contribution and whom does it benefit? These questions should be
used to lead the research deeper.

The authors need to dig deeper into the literature on meta-objects and
meta-classes. Brian Smith, Patti Maes are good starting points. Or the
Metaobject Protocol book by Kiczales.

*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*

Third reviewer's review:

     >>> Summary of the submission <<<

This pattern presents a schema for organising design patterns, based
on analysis of existing patterns and cybernetic theory.

     >>> Evaluation <<<

This paper makes a courageous attempt at a difficult problem - how to
organise and classify design patterns. The bulk of the paper (section
2.3) presents brief tour through many existing patterns.

Section 3 then discusses some general software engineering issues:
it's not clear what role this section plays in the argument being
developed by the paper.

Section 4 presents the new systematics for patterns. This starts with
an ad-hoc classification developed in section 2, and then includes a
number of other axes apparently based on the authors'cybernetic
theories.

I have three main issues with the classification. First, the
classification seems quite ad-hoc. The categories introduced in
section 2 seem arbitrary, and the paper presents no basis for the
"Human Thinking" in section 4.1 - for example, the paper needs
citations to the cognitive, philosophical, or psychological literature
to substantiate the basis of the classification.

Second, the classification treats patterns primarily as solution
structures. A pattern is more than just a solution and list of
participants (as the paper states in section 2). It should include the
problem the pattern solves, the important forces acting within the
problem, the dynamics of the solution (as well as its static
structure), the way the pattern resolves the problem forces or exposes
new forces (its benefits or liabilities). The presented
classification, however, focuses on the implementation techniques of
the patterns, neglecting these other important aspects.

Finally, the paper does not critically engage with prior existing
classifications of patterns, including Alexander's concept of a
pattern language (see e.g. Coplien's 1996 white paper on design
patterns, Alexander in IEEE SW 1999); Pree's metapatterns (AW, 1994);
Tichy's 1997 and Zimmer's (PLOPD1) classifications of patterns; more
academic analysis in terms of symmetry (Coplien; Zhao; Winn); language
dependence (OOPSLA 98); and semiotics (ECOOP 02).

Please note that this paper was not submitted in the correct
Springer-Verlag LNCS format for publication at ECOOP.


(To encourage accountability, I am signing all my reviews in 2005,
positive or negative. For the record, I am James Noble,
kjx@mcs.vuw.ac.nz)

*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*=--=*
