TODO: try to avoid a dependency on numpy in the test suite.

[dunshire.git] / TODO

diff --git a/TODO b/TODO
index b0b503bc4465b5048b83ffbe91b375ea73ac6e09..d0c0bf980f5bc1cc3ffba2dc1f0a1beb30d72147 100644 (file)
--- a/TODO
+++ b/TODO
@@ -1,15 +1,13 @@
-1. Add unit testing for crazier things like random invertible matrices.
+1. Implement the SPD cone.
 
-2. Try to eliminate the code in matrices.py.
+2. Make it work on a cartesian product of cones in the correct order.
 
-3. Make it work on a cartesian product of cones in the correct order.
+3. Make it work on a cartesian product of cones in the wrong order
+   (apply a permutation before/after).
 
-4. Make it work on a cartesian product of cones in the wrong order
-   (apply a perm utation before/after).
+4. Add random_game() and use it to replace all of the tests where
+   the cone is irrelevant. This should be done only after we are
+   feature complete and sure that everything works.
 
-5. Make sure we have the dimensions of the PSD cone correct.
-
-6. Come up with a fast heuristic (like making nu huge and taking e1 as
-   our point) that finds a primal feasible point.
-
-7. Should our one exception also spit out the game parameters?
+5. Remove the one use of numpy.matrix in a doctest, since that's
+   an annoying dependency to have just to run the test suite.