I am assuming (and, of course, I could be wrong) that the physical layout of the keys, as opposed to the assignment of functions to them, is fixed. Except that the ISO layout, as an alternative standard, would be a possible alternative. In that respect, I strongly favor the ANSI layout, because of the convenient location of the Enter key.
In another thread, I commented that the remote possibility (say, on some Linux systems) of the right and left Windows Shift keys serving different functions was enough to suggest
one change to the layout as depicted: instead of replacing the right Windows Shift key by the Fn key, replace the Windows Menu key by the Fn key. Since Windows Menu is not a
shift key, reaching it by an Fn-combination should be acceptable.
Other than that, I think the big issue would be to make the keyboard programmable. Some people would prefer a layout for the Fn-shifted keys like that of the HHKB, and others one like that of the upcoming MiniGuru. My own preference doesn't resemble either of those layouts.
When it comes to more radical changes:
For a product category between keyboards like yours, the HHKB, and the MiniGuru on the one hand, and a simple tenkeyless keyboard on the other, I've shown this illustration to the group:
to suggest that most of the keys outside the main typing area on a tenkeyless could be put in the traditional positions of the function keys, making it obvious how F1 through F12 would be reached with Fn, thus leading to a keyboard almost as compact as an HHKB, but a whole lot less intimidating (and therefore having a broader potential customer base).
Incidentally, the design as shown in the picture has the flaw of
not having an Fn key; instead, one uses the key that would be Fn-shifted to F8 to set up one's choice of Fn key. I don't anticipate a real commercial design being that radical.
