As a node evaluating a candidate delta, I want to score the delta based on my mempool content

[monsieurleberre]

In order to rank different candidate delta, as a node, I want to be able to

• produce a candidate delta from my own mempool
• use its hash to compare to the hashes of the candidates I receive
• if they are different, assess how much they differ (maybe using the MerkleTree) and attribute a score that decreases with increasing differences

[monsieurleberre]

very much related to #420

[monsieurleberre]

In the icebox for now, in the first implementation of POA, we simply rank the hashes by popularity, but are not able to use them to compare contents of the deltas.

[nshCore]

Let's update this ticket to see how it compares to our paper