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

E. Referee Comments (for author and editor)

Overall this an excellent presentation, and it seems to be entirely 
consistent with the practical orientation desired for this issue.

Having written asychronous distributed systems using PVM, I am however 
puzzled by the statement that "these systems [message passing systems 
like PVM and MPI] do not support the construction of asychronous systems 
as they restrict the application programmer to the Single/Same Program 
Multiple Data (SPMD) paradigm".  Perhaps a clarification is in order?

F. Presentation changes

It is sometimes difficult to visually distinguish figure captions
from adjacent text.  Also, the last lines of figure captions are
centered, rather than flush left which would seem more normal.

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

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

I would try to focus more on the software infrastructure of the PCE
rather than going into detail ovee the specific hardware
latency/badwidth Myrinet issues that seem to complicate the paper with
unnecessary detail. In the overview, I would like to know how more how
you see PCE as relating to Grid computing in terms of the PCE software
infrastructure if there is any overlap at all (and then describe areas
of overlap) or if you PCE really is separate from Grid computing, in
which case you may want to explain what makes PCE a Grid Computing
Environment per se.
What kind of security model does PCE use??
Good decription of the portable migratable sockets library.

As before, I think you have presented a substantial anmount of work in
the paper, but it needs to be more clearly organized around a central
thesis, beyond a description of the various research projects going on
at the Polder computer center. How does PCE envision Grid usage--
moreover how does PCE see any of its existing capabilities in terms of
Grid services.

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


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

This is interesting work, but the paper is to unfocused. 
It would be beter as three seperate papers. One on
cluster computing, another on networking, and one on computing environments.