Referee 1 *******************************************

E: Referee Comments (For Author and Editor)
 
  Summary: This paper describes on-going projects in grid computing, both in
  developing applications for a transatlantic test-bed and in developing toolkits
  under the DAMIEN project. The authors motivate the use of a message-passing
  model (specifically, MPI) and identify the need for debugging and performance
  analysis tools. The architecture of the DAMIEN system is presented and a
  collection of research and commercial tools are described. The focus of the
  paper is on their implementation of MPI (PACX-MPI) and an associated
  debugging and verification tool (MARMOT).

  This is a paper in the style of a "status report". The authors have been
   involved in grid computing for some time and describe an on-going
   project. It is hard to tell how much of the paper is about "new"
   things. Of the two specific tools, described on pages 5-7, one (PACX-MPI)
   seems to have been described earlier (c.f. reference #7) and used in
   several projects (c.f. references 4, 17, 21) while the other (the MARMOT
   debugger) appears to be new but not groundbreaking.

   One small question: PACX-MPI has been used on a number of projects.
   Have these projects employed some/most/all of the extensions listed
   in the bullet list on the top of page 6? Is there performance data about
   the effectiveness of the optimizations and data compression?

F: Presentation Changes

 1. The abstract is too vague. It should give a better preview of the contents of
     the paper.

 2. Page 2, 4th paragraph. The word "german" should be capitalized.
 3. Page 3, first line of text. The word "part" should be "parts".
 4. Page 3, 2nd line of text. "Message-Passing Standard" should be "Message-Passing
     Interface" or "Message-Passing Interface standard" or "message-passing standard
     MPI.
 5. Page 3, lines 4/5. "Supercomputers", "Computational" and "Grids" should not
     be capitalized.
 6. Page 3, 2nd paragraph. "figure 1" is probably better capitalized (Figure 1).
 7. Page 4, 2nd bullet on page. "applications" should be "applications". Also, the
     phrase "solve multi-disciplinary phenomena" might be better as "investigate
     multi-disciplinary phenomena" or "solve multi-disciplinary problems".
 8. Page 4, end of first full paragraph. " in semi-automatich" should be "in a semi-automatic".
 9. Page 5, 2nd line of text. "programms" should be "programs".
10. Page 6, last full paragraph. "Classical Debuggers" should be "Classical debuggers".
     Also, immediately after the reference [13, 22] there appears to be some missing
     words as the next sentence begins "approach is to ...".

Referee 2 *******************************************

E: Referee Comments (For Author and Editor) 
------------------------------

  The paper presents two tools that are definitely needed
for implementing MPI based application for use in distributed
an Grid environments, but (general criticism:) this is 
not all, that is necessary for a Grid Computing Environment 
(the topic of the CandCPandE special issue). The paper
would be more suitable for a Grid programming model special
issue.  

More detailed points and questions:

  The paper focuses on PACX-MPI and MARMOT, but both of these
software tools are not mentioned in the abstract.

  How is the 'integration' of the tools in the DAMIEN toolbox
done? A 'Integrated Software Development Environment' is more
than a couple of tools - relying on a common protocol or MPI - 
which can be invoked on the command line. Maybe you have
done more on the integration, then this should be expressed
in the paper, or the title of the paper is not feasable.

  The 'Project status and Future plans' section says, that
with the presented tools the use of a large number of 
applications is possible. But is there any application that
runs with PACX-MPI on a Grid infrastructure w/Globus or
UNICORE?  

  Also a comparison to other Grid-MPIs could be valuable.
Especially a comparison with MPICH-G2 would be interesting
to see (in terms of usage and performance).

  More technical: 
  How can PACX-MPI deal with firewalls? Does ist use a 
special port(-range)?

  Does MARMOT rely on PACX-MPI, or can it be used with any 
MPI implementation?


  The paper also doesn't say something about availability.
Is the software free? Can ist be used by others? Can it
be downloaded?

  Some pictures could also be added (e.g., something like a 
PACX-MPI architecture picture or a MARMOT screenshot).


F: Presentation Changes

  The paper should be reviewed relating to english punctuation.
It also has some grammatical errors.

Referee 3 *******************************************

E. Referee comments for author and editor:

The paper was pretty concise and to the point so their wasn't much to
critique. 

** Presentation - 4/5
     layout and flow of paper is good.
** Competence - 4/5
it showed how they used other grid-tools to create a
software development environment that made things easier for the HPC user.
** Contribution - 4/5
** Originality - 4/5
** Strengths - 4/5
 enables users to use the grid with more ease and better tools.