Horrors are written on the maps of Russia

[grin (Peter Gervai)]

@Minh_Nguyen

More importantly, usage of this built-in feature is tracked by the system, so the community has a much clearer picture of the database’s health based on good and bad edits. Tools like HDYC have only begun to track reverts, but they’re at a disadvantage because they can only guess based on diffs without any knowledge of the intent.

That is an important point: changeset metadata of reverts shall be formailsed to be easily parsed by machines.

Not just to count or follow it but often we have to undo the reverts and it is useful if the metadata contains the background (and I am intentionally vague here: I haven’t thought about whether there is any machine-parseable metadata required to undo reverts with the least damage).

(Which reminds me to mention that OSM reverts are extremely more risky and harder than Wikipedia: apart from really experienced editors (who can handle various kinds of conflicts) reversion is really not for the masses. Even I (with about 15 years of experience) often choose manual fix instead of reverting a changeset which had overlapping future edits (and trusting the tools to guess right what’s in and what’s not).)