?	Proposal #14

	I wouldn't consider this proposal for best proposal prize.

	I would give this proposal a 5/10 for overall clarity and organization. There were several places where the structure of a sentence needed improvement and there were some
grammatical/spelling errors (examples: the "sudo-embodiment" mentioned in the introduction, and frequent use of posessive "it's"). Also, the dataflows were difficult to
understand. However, the paper was decently written for the most part.

	The overall idea is original and interesting. As the researcher explains, it is a new idea that hasn't been previously tried. I really like it. 9/10.

	The idea of this proposal is that an Android phone will be used in conjunction with "Cellbot" (a mobile robot that can be controlled from a phone) technology to form a robot.
This robot will use visual input from the cell phone's camera in conjunction with SOMs to determine whether an object is hot or cold. However the interesting part to developmental
roboticists is that the robot will have no knowledge of this before training.

	This proposal meets most of the guidelines, giving a clear definition of what the researcher intends to do. As with many of the other projects, it doesn't quite meet the 10-15
page requirement, coming in at about 9 good solid pages.

	My favorite part: If the plan laid out in the proposal is followed, this will certainly make for an amazing project. Just think about it: being able to tell whether an object is
hot or cold, simply by looking at it? Better yet, this will be learned, not programmed. It fits in very nicely with Developmental Robotics. And just because of the sheer
"cool-ness" of the project, I can see this becoming a major Android app.

	Some sections of the project have been pushed aside in an "I'll deal with that later" sort of way, which leads me to believe that the researcher doesn't know for sure exactly
how these things will be accomplished. An example of this is the SOMs. The researcher explains what they are, but there are few technical details on how they will be implemented.
Another example is that a few libraries are described in passing such as the pHash library, which will be used if early experimentation with OpenCV "doesn't look promising".

	Also, I don't know if this project will be finished by the end of the semester or not. It seems very ambitious for someone who hasn't previously worked with OpenCV and has only
done a "little coding with ... [the Android SDK]".

	The dataflows, while decent psuedocode visualizations, are confusing and difficult to understand and could use polishing. In contrast, the mockups of the UI are very good, and
it looks as though the app has already had quite a bit of work put into it (either that or the researcher is very good at 'shopping an image).

	This project needs to be supported by relevant research papers and read more like a research proposal. Several links to projects/sites are included, but without the source and
authors it is difficult to judge the quality of the work. If you can't find anything specifically about Android developmental robotics, just cover some image processing and SOM
papers. It would add to the substance and quality of the proposal, and would probably help you in your project.

	My least favorite part of this project is the lack of clarity and the incompleteness of the paper. While the researcher does a nice job of explaining what he intends to do,
there are some standard details missing. The paper could benefit from page numbers, a Table of Contents, and a references list.

	In conclusion, the text of the paper was of medium quality. The idea behind those words, however, is an entirely different matter and if it is brought to life it will no doubt
be one of the best projects of the class. Good luck!