Currently, what we have are proposals which have priorities, with more specialised (e.g., block-diagonal TNT, bdmax) typically having higher priorities than less (e.g., TNT, bd), and hints hat can add or subtract from priorities.
However, as proposals and hints proliferate, we need to figure out some approach to making sure that usually the higher-scoring proposal is actually more efficient for the occasion.
I have an intuition that there might be some information-theoretic measurement of relative efficiency of a proposal, but I can't quite put my finger on it.
Currently, what we have are proposals which have priorities, with more specialised (e.g., block-diagonal TNT, bdmax) typically having higher priorities than less (e.g., TNT, bd), and hints hat can add or subtract from priorities.
However, as proposals and hints proliferate, we need to figure out some approach to making sure that usually the higher-scoring proposal is actually more efficient for the occasion.
I have an intuition that there might be some information-theoretic measurement of relative efficiency of a proposal, but I can't quite put my finger on it.