Subject: C436 final
Resent-Date: Fri, 12 Nov 1999 09:19:56 -0500
Resent-From: Geoffrey Fox <gcf@npac.syr.edu>
Resent-To: p_gcf@npac.syr.edu
Date: Tue, 9 Nov 1999 10:15:05 -0800 (PST)
From: Chi Chao Chang <chichao@yahoo-inc.com>
To: gcf@npac.syr.edu
CC: chichao@yahoo-inc.com

Geoffrey,

Enclosed is the final version of the paper.  I apologize for the delay
and hope it is not too late for you.

I was able to address most of the suggestions raised by the reviewers.
Below is a synopsis (per reviewer):

Reviewer 1:

- addressed all typos and errors pointed out

- unable to compress figures 2-3, 5-6 and 11-12 into 3 figures.  The
  reason being that there is already too much data to view; if
  compressed, it becomes even harder for the reader to follow.  Given
  that we have no space constraints, we decided to keep them as is.

Reviewer 2:

- we addressed a main concern about the lack of implementation details
  in Javia-II by adding a 3-paragraph section on it.  We think it
  provides enough details for a good understanding.  I also refer the
  reader to my dissertation if they would like to know more about it.

- we tried to improve the presentation of results concern as much as
  possible.  We made a clear distinction between Marmot numbers and
  those obtained by Sun's JVM running JDK1.2.  We did not change the
  graph legends since we feel that they are clear enough and only one
  reviewer complained about it.

Reviewer 3:

- addressed all the typos and errors pointed out

- we explicitly stated that Javia-II and ViBuffers cannot be
  implemented on an off-the-shelf JVM, addressing the reviewers
  concern about whether it could be done on Sun's JVM.

- we did not include additional explanation on garbage collection
  because it goes beyond the scope of the paper.

Thanks.

-- Chi-Chao

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