The organization of this paper is very good: background in qos, framework 
for a qos-aware operating system, comparison of existing os's, summary.
However, my one major complaint is that the sections describing the 
framework comprise a disproportionately large part of the paper.
I think the paper could have been improved if the framework was 
described in less detail.  

The summary section was a little lacking.  In particular, I found it
quite striking that *none* of the os's offered any of the I/O system
features described in the framework.  Although the authors mention this
observation, no possible explanations are given.  How do we know that 
the lack of these features is truly a weakness?  Do the os's make up for 
it in any way?  Does this say anything about the assumptions presented in
the framework?

The readability of this paper is somewhat poor, which is why I gave only
average scores for quality.  Although it's partly a matter of 
preference, I still think bullet-ed lists are used too liberally in this
paper, particularly in the framework-related sections.  Additionally, I
think the paper has some formatting problems which hindered readability: 
Be sure to indent paragraphs correctly, and try to incorporate the 
figures in the text (figures 1,2).  

Overall, I gave the paper above average marks because I thought it 
presented a novel idea in a relatively clear fashion.  The framework, 
although lengthy, provided good direction for the reader.  I also 
thought the summary section presented some good observations, particularly
the idea that existing page-replacement algorithms may be unsuitable
for managing multimedia streams.  These observations though should be 
developed more, and could possibly be used as future areas of research. 

======================================================================

Technically, the paper is fairly dense but very hard to read. I read it 
twice and still didnt' extract much lasting content from it. The framework 
for characterizing MM-aware OSs was not well explained and doesn't 
necessarily map well to all areas of OSs IMHO. The authors used many terms 
without defining or explaining them, instead relying on the user to 
understand them already or to pick up the context and meaning implicitly, 
which shows that the target audience is not well understood. No MM or RT 
knowledge should be assumed in a general OS grad-level class. MM-aware OSs 
are a nifty research area since MM is a hot topic but I didn't really 
capture much from the paper after reading it about why this is important, 
how their framework can be used to judge different MM OSs, why these 
features are even important in MM OSs.
Authors didn't follow directions: hard limit of 10 pages was exceeded. 
Paper was 15 pages long. Odd word spacing in paper was hard to read, as was 
having poorly delimited paragraphs (no indentations at beginning of 
paragraph). Two pages of figures were not incorporated into the text and 
were not explained well; flipping from the paper to the figures was 
distracting. Lots of poor grammar and illegible sentence construction made 
it difficult to extract content from the paper. The authors should 
proof-read for good grammar (or have someone else do it) before submission. 
Section 3 doesn't flow well with the rest of the paper.

======================================================================

This was an interesting topic, but the repetitious layout of
the paper made for tedious reading.  The diagrams were helpful
in making the material understandable.

======================================================================
The topic was interesting and I think that it is a reasonably "hot"
area of OSs right now.  Real-time MM is definitely something that
people are interested.

I do not have much experience with MM but the observations for the
most part seamed to be somewhat novel.  (this could be due to my lack
of MM experience).

The quality of the observations was good.  In the Memory Management
section it appeared that there was not enough information to make a
meaningful comparison.  Chorus was the only system that was critiqued.

Overall The paper had some good points, but the presentation was a bit
awkward.  Section 4 seamed very redundant; the ideas could have been
expressed in a couple of paragraphs instead of 2 1/2 pages.  The
conclusion was a bit unclear; there were some nice charts, but there
was no explanation of how to interpret the charts, and there was no
key to explain what a check meant compared to and 'X'.  The division
of the classes was clearly defined.  The classifications were good.
The systems where clearly compared against the criteria presented in
sections three and four.  Although sections three and four did not have
much flow to them.  They were entirely built of one paragraph
subsections that changed topic.  The paper had some good classes and
classifications, but the lack of continuity made it difficult to
absorb. I thought that there was some good ideas presented and with a
little more thought to the organization this can be an excellent
paper.

======================================================================

The topic is really interesting and important. We need to build an OS that
meets the QoS requirements for multimedia applications. The font is not 12
point and the paper is 15 long with the figures, although it would probably
meet the conference requirement (10 pages) if the right size of font was
used. The appearance of the paper doesn't meet the conference requirments.
There are also some grammar mistakes, which indicates luck of proofreading.
The authors provide the framework for a QoS-Aware OS as if it was then that
came up with these requirements. The paper is well structured and it gives
suffiecient information about the topic. The 4 Multimedia OS that are
presented are being compared to each other but also to the ideal framework.
The future research directions shouldn't be under the section "Summary and
Discussion", because the summary and the future directions section should be
seperate. In the conclusion section there is not really a conclusion, but
rather a summary of what was presented in the paper.

======================================================================

The paper is well structured and easy to read.  The analysis 
which the authors provide is clear and concise.  I especially
like the table summary in section 6, "summary and discussion",
for it gives a clear picture of the comparison parameters and
results.   

======================================================================