It's really nice to see this topic being brought up again. Also I had never seen that thread about my screens A1ex

.
With all the development ML has been through, and only broadening it's possible scope with all the possibilities still ahead (ahem DIGIC Registers), this effort really should be about first and foremost making the UI flexible, robust and easily expandable, and only second about color, contrast, gradients and icons.
I made my first versions bases on an idea of making the functions more visible, which did result in some extra clutter. Can't really say it was a very involved effort or that there was much feedback

. Although it did please me to see my integrated help suggestion making it in and improved upon.
I believe the best way to go would be to create a core UI development group with a few people. This way we can work faster, more focused, and without posting every little detail (but keep info flowing and gathering feedback from the community).
With this setup the best way would be to work with wireframes of the UI, exploring well the menu mechanics and flexibility of the system. This should include the menus, different camera res, even help system (moving away from the IMG based one).
After this the focus could then be on color, technical issues with fonts, themes (low/high contrast at least) and polishing.
Let me know what you think of this, I will contribute to the effort no matter what the outcome of the development may be - Now let me just find and dust of those old files...
Furthermore I have to thank all the ML developers, testers and contributors for the amazing tool we now have.
Although I really have used it little in the past 8 months (proud father of a very energetic 4 month old girl) I'm eager to see it developing further and improving in every step of the way. As much as I am to exploring it further and producing some amazing images with it.
PS: I really see no issue with the menu being different from the standard Canon menus, for me it's all about function and we don't need any kind of artificial limitation in trying to look similar or even having that as a start point.
I would also make the case that the menu changes a bit in every camera ML supports, particularly the older ones (can we get a few screens from each as reference!).
Furthermore, I don't really think that the fact of it being different distracts in any way, you just know you are in ML

(where the good stuff happens).