Subject: submission CPE java special issue
Resent-Date: Fri, 03 Sep 1999 14:46:54 -0400
Resent-From: Geoffrey Fox <gcf@npac.syr.edu>
Resent-To: p_gcf@npac.syr.edu
Date: Wed, 18 Aug 1999 09:25:38 +0200
From: Michael Philippsen <phlipp@ira.uka.de>
To: gcf@npac.syr.edu
CC: phlipp@ira.uka.de

Geoffrey:

here is our paper on complex number for Java that has been
accepted for ISCOPE. I include the author notification mail
with the reviewer's comments. Our final version will have
what Referee 2 suggests.

I hope that it is not too late for the paper to be considered
for the special issue.

Please let me know if you have any problems printing it (or
converting it to PDF).

Regards,

        Michael.

> From: MaryDell Tholburn <marydell@lanl.gov>
> Subject: ISCOPE '99 Notification of Paper #38
> Cc: matsu@is.titech.ac.jp, marydell@lanl.gov
>
> Drs. Michael Philippse and, Edwin Guenthner,
>
> We are pleased to inform you that your paper
> "Complex numbers of Java"  #38 has been accepted by the ISCOPE'99
> Program Committee for publication in the conference Proceedings as
> a full paper (12 pages), and for presentation at the conference in
> San Francisco, December 99.
>
> The  overall acceptance rate was about 34% for full paper.
>
> Please  start revising your paper according to the referee
> reports. You will received a mail from Rod Oldehoeft (Proceedings Editor,
> rro@acl.lanl.gov) to help you to create the final version of your paper
> for inclusion in the Proceedings. The camera-ready deadline is
> September 8, 1999.   This is a firm deadline no extentions will be
granted.
>
> Referee Report:1
> A compiler or preprocessor that supports a dialect of Java to which
> complex numbers have been added and produces either Java source code
> or byte codes. The necessary translation details are exhibited at a
> good level of detail so that one can understand the translation and
> optimization issues.
>
> This is an outstanding paper. It is interesting and topical; it has
> some novel ideas; it has actually been implemented and measured; and
> it describes what was done in enough detail that I think I could
> replicate and verify the work.
>
> Referee Report:2
> This is a very well written paper that addresses the important topic
> of handling complex arithmetic useful for pursuing scientific programming
> in Java. The work is thorough and the implications of decision making in
the
> representation of the complex type are explained. In general, this is a
fine
> contribution that suggests alternatives to a class-based approach for
complex
> number representations.
>
> At times the paper may be a little too detailed for a general audience,
> but I feel that these details are necessary for scientists interested in
> this topic. I could not find a discussion of how complex conjugates
> are represented/calculated, so perhaps this could be introduced in
> section 4.2 in a later revision. This is an important function for FFTs,
> of course....
>
> I would like to see an extended (or more complete) disussion on the
> choice of using two arrays to represent a single complex array. I would
> suspect that the perform impact of this representation regarding array
> accesses and initialization could be important, yet this topic has been
> omitted in the draft. The question of how pairs of double-values may be
> stored should be justified, since there is no immediate reason to guess
> that they may not be stored in adjacent positions if allocated and
initialized
> in this manner - even when optimizations are applied. Some languages
provide
> constructs to even ensure that this will be so.
>
> Finally, the benchmark results are impressive and justify the author's
> efforts. This is a fine contribution.
>
> Referee Report: 3
> Paper presents an approach for including complex numbers in Java and
> presents experimental results to back up the claims.
>
> Referee Report: 4
> Paper presents issues of adding Complex types to Java in an efficient
> manner. It provides of the transformations required and justifies their
> success with numerous experiments. The paper is focused, brief, and
reaches
> a clear conclusion. Well done.
>

    ---------------------------------------------------------------------
                        Name: complexe.ps.gz
   complexe.ps.gz       Type: Postscript Document (application/postscript)
                    Encoding: x-uuencode
                 Description: complexe.ps.gz