The part I and part II of this paper are available if you haven't read them. This is to summarise and mention what has being done.
To summarise, the 3D Face algorithm is:
the spread of data unto the congruent convex regular polygon faces of the platonic solid
1. A public key which determines face rotation.
2. A private key determining face layering.
What I have failed to define are quite a number. If really I'm up to something here, the specific mathematical functions or operations I continually refer to would be defined.
These would be areas such as:
1. Determination of the base face (face 0).
2. Operation on fixed length data from datastream before write.
3. Storage of data on face such as not to overflow (see step 3 of part I)
4. Operations to determine face rotation.
5. Specifics on storing data on a face in layers.
And other little bits which may come up in the implementation of the above.