I somehow fail to see why the example in the article needs the hash to have some special property of homomophicity. When you represent the overall hash of your dataset as a sum of individual item hashes it trivially follows that change of one item means that you substract the hash of original version and add the hash of new version.
Or am I missing something? (apart from the somewhat obvious security implications of doing such a thing in the first place)
Or am I missing something? (apart from the somewhat obvious security implications of doing such a thing in the first place)