@@ -1337,4 +1337,19 @@ where $I_\Gamma$ has columns $\vec{e}_\gamma$ for $\gamma \in \Gamma$.
We could also calculate numerators and denominators separately and combine them as needed.
\subsubsection{Notes on current implementation}
Current implementation (IntersectSimplices v1, Unfold v1, Orphan v2 and IntersectHyperplane v1) is in working order and compatible with ProjectionMatrix3D.
It is not clear if the unfolding process has any overlaps / is in any way efficient.
Certainly the orphaning could be done more smoothly but we're stuck on certain graph problems.
Also missing is the check of signs from previous generation of intersections indicating signs of next generation.
If two intersections in a given generation have the same sign in a given coordinate then the sign of any intersection in the next generation with the same vertices has the same sign in that coordinate.
As explained somewhere above if there are fewer than $m$ intersections between $m$ hyperplanes of $Y$ and an $m$--face of $X$ then the orientation of the face is predetermined by other intersections.
The size of the matrix $D$ which holds the values of the various determinants is much larger than it needs to be.
Its current size is determined by the easy way of translating between the binary strings for $J$ and $\Gamma$ and integer coordinates (literally convert binary string to integer).
This leaves a lot of blank spaces in both directions.
It is almost certainly preferable to a) come up with a new bijection between the binary strings and integers and b) use sparse matrix for $D$ (but still global?).