The problem with writing a new parser in Python, is that the implementations would probably end up drifting. New features and bug fixes would need to be built twice, and you would also find bugs in one implementation that are not present in the other.

I'd be more interested in re-writing in JS, as then it could be used for both desktop (using say V8 or Java/Rhino) and for the web. The compiler is also about 10,000 lines, which excludes the Quby libraries (String, Array, Hash, Function, Image, Controlls, Sound, etc), and I'd rather re-use as much as that as I could for any desktop port. The only bit that really needs re-building is the grammar, where the structure is formed.

Is there a command for checking the color of a certian pixel? That would be useful, if there isn't one already.

There is, although indirectly. You can call the 'getScreenshot' function to get a copy of the screen as an image. On the Image object you can then call the 'getPixel' methods, or similar ones such as 'getPixelRed', to get specific pixels from the canvas.

and the paint editor I built for the site, has just been made a Chrome Experiment. Yay!

Nice paint program