Ensure that "overflow" classifications, post-tranche, are handled correctly

[bogden1]

By overflow classifications, I mean that we have run a classification through the scripts and it has been delivered to the project team for manual review, and that a volunteer has then submitted an extra classification.

Although I'm loathe to drop any volunteer work, I think that the only rational thing to do in this case is to drop the classification. We could detect these cases, see whether the new classification changes anything, and flag for re-review, but that is a lot of complexity and I do not believe that we have time to do this work.

We probably should discourage volunteers from re-classifying retired subjects. It is probably better to comment on the Talk pages if the volunteer is concerned about something.

So this task is just to check what happens to such classifications in the current tranche implementation, ensuring that we drop them rather than add them to a new tranche -- ideally producing a report of any such cases.

[bogden1]

My current thinking is that we will just allow phases to complete and not worry about tranche support, so dropping priority on this.