answer contacts can unsubscribe anyone

[Eliah Kagan]

I am an answer contact for several Ubuntu packages, and I have noticed that I have the power to unsubscribe other users from questions, even if I did not myself subscribe them to the question.

It seems to me that the benefit of allowing this, if any, would probably not outweigh the detrimental effects of having users accidentally unsubscribe each other. Rejecting a question is a more complex action than unsubscribing a user, but even if an answer contact did accidentally (or otherwise wrongly) reject a question, its subscribers would receive notifications and could then act to get the mistake corrected. If an answer contact accidentally (or otherwise wrongly) unsubscribes a subscriber, it seems that this systematic mechanism for getting it corrected does not exist.

I asked about this, and Gavin Panella expressed the view that this was odd behavior which might be wrong. So I'm filing this bug.

[Robert Collins (lifeless)]

Users get subscribed and cannot figure out how to unsubscribe. E.g.:
 - an app crashes
 - apport asks them to file a bug
 - the bug is marked a dupe of another thing
 - thats converted into a question

And the poor user has been carried along with no idea of whats going on, and they start complaining (via email reply-to): allowing responsible folk to unsubscribe them allows this to be fixed.

We possibly should notify folk unsubscribed by a third party. Possibly.

[Robert Collins (lifeless)]

However, as this was filed out of concern rather than having a specific problem, I'm going to mark it invalid - this is a deliberate situation.

[Eliah Kagan (degeneracypressure)]

I agree with the status change to Invalid. If at some point I feel overwhelmed with the belief that the current behavior should be changed, I'll file a new bug with a clear indication of what I think it should be changed to and with a clear argument backing up my viewpoint. Right now, the explanation in https://bugs.launchpad.net/launchpad/+bug/880960/comments/1 makes perfect sense to me. It had not occurred to me that a user might not know *what* it was that they needed to unsubscribe from.