This dispute is submitted by authors of the paper.

We'd like to point to a number of errors in the review of our artifact.

data.txt
  - "student5: the downloaded file just consists of bunch of word documents"

    This is simply not true. Possibly this comment referred to a different
    paper? We do not use MS Word. The tarball we provide consists of source
    code, plaintext readme files, and some binary files (e.g. JARs etc).

build_notes.txt
  - "Btrace: can be downloaded from http://kenai.com/projects/btrace"

    This is irrelevant. The artifact contains both the original version of
    btrace and our modified version.  There is no need to download btrace.

  - "I installed mercurial and then I tried to use the command $hg clone
    http://jikesrvm.hg.sourceforge.net:8000/hgroot/jikesrvm/jikesrvm I got
    abort: error: connection timed out, I tried 3 times and got the same
    error"

    While the mercurial repository of Jikes may have been down when the
    reviewer tried, it is working now.  Moreover, to mitigate similar
    problems we included Jikes source code into the artifact, in the very
    same directory where the readme with the mercurial URL was mentioned. 
    There was no need to download Jikes in the first place.

    It seems that the review concluded that because of the failure to
    download Jikes, the artifact cannot be built (see "jikes failed to
    build" in data.txt). This is in our view an incorrect reasoning. (*)

misclass.txt
  - "The code for both tools is available and compilable, however the
     modifications required to reproduce the paper results do *not* appear
     to be available."

    The modifications of both btrace and Jikes RVM are available in the
    artifact.  See bytecode-instrumentation/btrace for modified btrace and
    jikes-extension/jikesrvm for the modified Jikes RVM.

    Now, we indeed tried to build our own artifact again now when told about
    this study. The modified Jikes RVM is actually not compilable without
    some effort (*), so even the statement in the review, this time in our
    favor, is incorrect.

(*) We found that the jikes extension is actually not compilable without
some effort (e.g.  one has to include the DaCapo JARs into the build
directory).  As we are authors we are conflicted to judge whether the effort
needed is reasonable or not.  It does not seem that the reviewers in the
study actually tried to build the modified Jikes RVM code.

We will fix our artifact to make building of our Jikes RVM modification
easier. However, we believe that only people who have some experience
working with Jikes RVM could get any benefit from using or building on our
tool, anyway. From the review it does not seem obvious that this was the
case for the reviewers.