Hello,
Thanks a lot for Cadabra. This is very useful!
But I found strange behavior of the algorithm meld.
First, a bug related to prefactor.
Consider:
ex:= B^{i} B^{j} S^{m n} D_{i j} D_{m n} + B^{i} B^{j} S^{m n} D_{m n} D_{i j};
S^{i j}::Symmetric;
D_{i j}::Symmetric;
And then if I evaluate meld for the ex, I get prefactor 3/2 instead of 2.
Moreover, if I do not specify the symmetries of S^{i j} and D_{i j} separately, then the algorithm works correctly.
Second, an incorrect interpretation of an index convolution:
\partial{#}::PartialDerivative.
ex1:= G^{m j} \partial_{i}{A} \partial_{m}{F_{j}} + G^{m j} \partial_{m}{A} \partial_{i}{F_{j}};
ex2:= D^{m n} \partial_{i}{C} \partial_{m}{Q_{n j}} + D^{m n} \partial_{m}{C} \partial_{i}{Q_{n j}};
meld(ex1);
meld(ex2);
Meld interprets expressions as two identical terms in both cases, although I did not specify any symmetries of objects.
Thanks,
Maksim