X-Git-Url: http://gitweb.michael.orlitzky.com/?a=blobdiff_plain;f=mjo%2Feja%2FTODO;h=f27df9cdc14c6e94cabd6a067285bc354ecd3cbe;hb=f98ab4d7afa92a853e7ddc75cdac803d2da4fcb9;hp=1407ebdbe83bafab1ef79d881d25d2cf4912e18e;hpb=ee9ac102b8b392793466c13039a6e50b1e3c4c01;p=sage.d.git

diff --git a/mjo/eja/TODO b/mjo/eja/TODO
index 1407ebd..f27df9c 100644
--- a/mjo/eja/TODO
+++ b/mjo/eja/TODO
@@ -1,11 +1,9 @@
-1. Finish CartesianProductEJA: add to_matrix(), random_instance(),...
-   methods. I guess we should create a separate class hierarchy for
-   Cartesian products of RationalBasisEJA? That way we get fast
-   charpoly and random_instance() defined...
+1. Add cartesian products to random_eja().
 
 2. Add references and start citing them.
 
-3. Implement the octonion simple EJA.
+3. Implement the octonion simple EJA. We don't actually need octonions
+   for this to work, only their real embedding (some 8x8 monstrosity).
 
 4. Pre-cache charpoly for some small algebras?
 
@@ -17,3 +15,8 @@ sage: a0 = (1/4)*X[4]**2*X[6]**2 - (1/2)*X[2]*X[5]*X[6]**2 - (1/2)*X[3]*X[4]*X[6
 5. Profile the construction of "large" matrix algebras (like the
    15-dimensional QuaternionHermitianAlgebra(3)) to find out why
    they're so slow.
+
+6. Instead of storing a basis multiplication matrix, just make
+   product_on_basis() a cached method and manually cache its
+   entries. The cython cached method lookup should be faster than a
+   python-based matrix lookup anyway.