to take into account the size of revision
I think for example of the first author of an article who can had only 1 revision, but has produced 80% of article bytes.
Alexander Furnas commented
Something along the size of revision would seem to be useful - however - what we really want is *importance* or *significance* of revision. This is harder to quantify given that size is only one aspect of significance. Perhaps some sort of crowdsourced rating system or tagging of significance could be implemented - people could flag key edits that are of note for various reasons. Potentially this could be used as a form of oversite to look for systemic abuse among high level editors. Perhaps some sort of slashdot moderation + meta moderation thing could filter out incorrect tagging of abusive edits. Just some thoughts.