RELEASE FORM

Please print out and complete this form, sign and fax to (714)821-4010,
Attention:  Manuscript Assistant, or mail to Manuscript Assistant, IEEE
Transactions on Software Engineering, 10662 Los Vaqueros Circle, Los
Alamitos, CA 90720 USA

I certify that except as noted below

Manuscript #110422, "A Comparative Study of Clustering Algorithms and
Abstract Representations for Software Remodularization"  Corresponding
Author: Timothy Lethbridge

has not been published elsewhere nor is it currently under review for any
other publication.  I will notify the Editor-in-Chief of TSE if I submit
this paper elsewhere prior to a final decision being made on publication
in the IEEE Transactions on SE. 


__________________________________                 __________________
Signature of Corresponding Author                                Date


*Exception(s): Please attach a copy of the paper that has been published
or submitted elsewhere as well as a short description of the differences
between that paper and the one submitted to TSE.  SEND ONE COPY TO THE
EDITOR ASSIGNED TO YOUR PAPER: 

The following paper was written prior to our TSE submisstion and covers a
few of the same experiments and results. The TSE submission is based in
part on this paper, but has been very significantly revised and expanded.

Anquetil, N., Fourrier, S. and Lethbridge, T.C. (1999), "Experiments with
Hierarchical Clustering Algorithms as Software Remodularization Methods",
Working Conference on Reverse Engineering, 1999, in press

The differences are twofold:  First, when we wrote the TSE paper we
significantly expanded on some issues already present in the WCRE paper
(to improve readability or scientific value);  second, we _introduced_ new
issues to better cover the whole subject.

Detailed differences are:

(_introduction_) Dramatically extended the discussion on different
  descriptive features.  This material was hardly mentioned in the WCRE
  paper whereas it is the most important point in the TSE paper.

(_modification_) New experiments: Based on the initial results
  presented in WCRE paper, we re-ran all the experiments with slightly
  different parameters.  This resulted in more meaningfull and easier to
  compare results.

(_introduction_) New experiments: We added completely new experiments
  (Note: correlation between descriptive features) to highlight some new
  points that were not discussed in the WCRE paper.

(_introduction_ and _modification_) Extended discussion on related
  literature to better cover various issues as well as to cover new
  issues.

(_modification_) Re-wrote significant parts of the WCRE paper to try to
  clear up some points that were obscure.