From: Michael Orlitzky <michael@orlitzky.com>
Date: Sun, 21 Feb 2021 22:29:07 +0000 (-0500)
Subject: eja: update the TODO.
X-Git-Url: http://gitweb.michael.orlitzky.com/?p=sage.d.git;a=commitdiff_plain;h=cc309b502850fd99abc742ee4ee4e015312f65ac

eja: update the TODO.
---

diff --git a/mjo/eja/TODO b/mjo/eja/TODO
index 25ef99d..8f9c17f 100644
--- a/mjo/eja/TODO
+++ b/mjo/eja/TODO
@@ -1,8 +1,11 @@
-1. Finish DirectSumEJA: add to_matrix(), random_instance(),
-   one()... methods. Make it subclass RationalBasisEuclideanJordanAlgebra.
-   This is not a general direct sum / cartesian product implementation,
-   it's used only with the other rationalbasis algebras (to make non-
-   simple EJAs out of the simple ones).
+1. Finish CartesisnProductEJA: add to_matrix(), random_instance(),
+   one()... methods. This will require rethinking what a "matrix
+   representation" and "matrix space" means for a cartesian product
+   algebra. Do we want our matrix basis to consist of ordered pairs
+   (or triples, or...)? Should the matrix_space() of the algebra
+   be the cartesian product of the factors' matrix spaces? Can
+   the FDEJA initializer be made to work on tuples, or will it
+   need to be overridden?
 
 2. Add references and start citing them.
 
@@ -20,4 +23,4 @@ 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
 
 6. Profile the construction of "large" matrix algebras (like the
    15-dimensional QuaternionHermitianAlgebra(3)) to find out why
-   they're so slow.
\ No newline at end of file
+   they're so slow.