2010 Progress Report
Rubik's Stellated Dodecahedron

Here's my letter to Melinda Green: 
  At

 	http://www.cs.umb.edu/~eb/rubik/

  you'll find some information about a rubik-like puzzle I've imagined,
  based on the small stellated dodecahedron. 

  There's also a (flawed) Java implementation (applet and jar) of a
  virtual version - I have ideas but no good ones about how to build a
  physical incarnation.

  Can I interest you in coding up the virtual version so that it works
  as it should? In principle I could do that myself. In fact I'll never
  get around to it.

  Ethan Bolker
and (part of) her answer: 
  Hello Ethan,

  Your puzzle is interesting ...

  I'd be extremely unlikely to help you with your software but I encourage 
  you to join our 4D cubing Yahoo group and ask politely for suggestions and
  possible help. Just be aware that the  members are mostly focused on
  higher-dimensional puzzles but will certainly also be interested in 3D puzzles.

  My opinion is that this puzzle is best suited to a physical
  implementation and it does not appear too difficult. You could begin
  by simply augmenting a Megaminx with pyramidal face tips. ...

Good luck!
-Melinda
I looked at the Megaminx, and at Alexander's star to see if either suggested a way to build a physical incarnation. I think not. What distinguishes my puzzle from these and from others I've see is that each of the (60) facets is independent of the others. There is no equivalent of the Rubik's cube corner (or edge) cubies, each of which carries three (or two) squares along with it when it moves.

So I think a virtual version is required, at least to begin with. I wonder if the framework PuzzleDescription.java on the magiccube4d Google code site is generic enough to implement my stellated dodecahedron. I note that it does provide for the three dimensional Schlafli symbol {5,3}, although not for {5/2, 5}.

I hope the challenge is interesting enough to engage some of the 4D cubing developers even though the puzzle is only 3D.