bgirardot
commented
7 years ago @majkaz This is an interesting issue and one I had not thought of. I am not sure what we can do to keep track of or check for editing outside of the tasking manager as it relates to validation. It might be training related for validators so they can be aware of it.
This connects to the evaluating of mappers - sometimes a task gets invalidated because of later edits and not because its state when marked as complete. It might be useful to distinguish this case (and the self-invalidated cases when wrong button got pressed) from invalidating because of incorrect work.
This will happen more often if the validation takes a long time - I have done it myself, when there was so much wrong it took almost complete reworking and I didn't have time to do it at the moment. The task was at the time completed for 4 months and several edits came 2 months later. The changes didn't correspond exactly with the tasks, meaning it probably came from outside of task manager.
But it has very recently happened on task I completed 6 days ago - there was a mess of additional changes from OSM outside of the task manager later, visible only in OSM history. With the number of mappers growing, the map is changing very quickly now, even in comparison to a year ago, for example. I have seen changes as soon as few hours after a task was marked done through task manager (visible only by the unlocking and relocking corresponding with the author of the changes, without any comments in the task or in the changeset) as well as changes completely outside of the task manager, often without any comments as well.