X-Git-Url: https://gitweb.michael.orlitzky.com/?a=blobdiff_plain;f=TODO;h=ab87751f3168ceecdf739128a222dd766828c377;hb=920c93e69916d70c293c8ab0caccb0b4d8ffb073;hp=4dee0ae8fb2e832b96c845ebaef1e82f1eb89337;hpb=b1d06cfefc83b4db9d5592f7424d4ef67b9806e7;p=dunshire.git

diff --git a/TODO b/TODO
index 4dee0ae..ab87751 100644
--- a/TODO
+++ b/TODO
@@ -1,29 +1,23 @@
-1. Add doctests for simple examples like the ones in Dr. Gowda's paper
-   and the identity operator.
+1. Add unit testing for crazier things like random invertible matrices.
 
-2. Add unit testing for crazier things like random invertible matrices.
+2. Add real docstrings everywhere.
 
-3. Test that the primal/dual optimal values always agree (this implies
-   that we always get a solution).
+3. Try to eliminate the code in matrices.py.
 
-4. Run the tests with make test.
+4. Make it work on a cartesian product of cones in the correct order.
 
-5. Use pylint or whatever to perform static analysis.
-
-6. Add real docstrings everywhere.
-
-7. Try to eliminate the code in matrices.py.
-
-8. Make it work on a cartesian product of cones in the correct order.
-
-9. Make it work on a cartesian product of cones in the wrong order
+5. Make it work on a cartesian product of cones in the wrong order
    (apply a perm utation before/after).
 
-10. Add (strict) cone containment tests to sanity check e1,e2.
+6. Rename all of my variables so that they don't conflict with CVXOPT.
+   Maybe x -> xi and y -> gamma in my paper, if that works out.
+
+7. Make sure we have the dimensions of the PSD cone correct.
 
-11. Rename all of my variables so that they don't conflict with CVXOPT.
-    Maybe x -> xi and y -> gamma in my paper, if that works out.
+8. Use a positive tolerance when comparing floating point numbers.
 
-12. Make sure we have the dimensions of the PSD cone correct.
+9. Come up with a fast heuristic (like making nu huge and taking e1 as
+   our point) that finds a primal feasible point.
 
-13. Use a positive tolerance when comparing floating point numbers.
+10. Ensure that we only compute eigenvalues of *symmetric* matrices
+    (so that the eigenvalues are real).