Sorry, but this is good on answers.

Following a post that does not actually have future activity is not really a problem, is it? Similarly having a large list of followed posts does not seem to be problematic either. The user profile contains this list, but that is not really needed imo, the nice thing about following is that you don't have to look at the list because you get a notification when something happens to any of the posts on the list.

@Marijn see the edit.

@VScode_fanboy Historical locks can only be imposed on questions.

If you don't want to get notifications then you obviously shouldn't follow posts. Whole point of this FR is to get notification if something changes, no matter how unlikely it may be. Bookmarks don't send notifications.

In the edit you mention that a large following list can flood your inbox, which is a fair point in general, but it doesn't really apply to the current question I think - if it was possible to put locked posts on the followed list you do that because you want to receive updates on those questions, so if enlarging the lists with these posts means you are now getting (more) flooded then this is exactly what you want (because the extra flood are updates you are specifically interested in).

But maybe I am a bit biased because my main site is relatively small, and most questions only generate very little activity after a while.

@Marijn I only follow the strictly essential technical stuff I'm interested in, and even there the notifications are too much (and managing any list is a pain). As for Dalija's objection I'd have to repeat a previous question (on a now deleted long comment chain) how many posts have we seen where historical locks get reversed? None? Does that ever happen? In any case I'm repeating myself at this point.

@ResistanceIsFutile I wanted to address the narrow scope of the question, but you wouldn't believe the amount of besides-the-point arguments (some now deleted) raised in the comments that change the subject.

I was fine with your first version - it directly addresses FR, the other points are irrelevant as people can freely follow or unfollow posts. Or should I say, I don't agree with your first version either.

@ResistanceIsFutile (jokingly and seriously) and if a SEDE query proved no historical lock was ever reversed, would that suffice? Or would we still need an FR proved to have no practical use just for the sake of having it? (You know others would argue a functionality having no use should be removed - as to "not confuse people", "bloat the feature set" and "needlessly consume dev time".) But whatever the case, I neither agree nor disagree, I write a righteous post because it deserves to be written.

I found some of these from SO via SEDE: stackoverflow.com/posts/1964839/timeline, stackoverflow.com/posts/38060436/timeline, stackoverflow.com/posts/9563675/timeline, stackoverflow.com/posts/246215/timeline. Many times a historical lock is reversed only to apply a different lock, but for some posts the lock is removed entirely and the question is left open. Indeed this is uncommon (for SO at least), my query returned a few hundred results, most of which were false positive (unlocked only to be closed and locked again).

But indeed as @ResistanceIsFutile mentions this is probably an easy modification to make, it would even simplify the code for following because you don't need to check for locks, so it is win-win :)

The query is data.stackexchange.com/stackoverflow/query/1511146/… but I'm not happy with it because of the false positives, unfortunately my SQL is very rusty so I'm not sure how to solve it easily.

@Marijn according to that query only 1 thread was closed in 2021, 5 in 2020, etc... Is that accurate? But remove the duplicates and filter only for questions that ended up reopened (that's the focus of this FR). (Currently the query also gives results for content dispute locks, comment locks, etc... Maybe those controversial posts should be debated separately since arguably they're better left without notifications.)

The query is for historical locks only (post notice id=22, so the query does not show content dispute etc) that have an unlock event (historytypeid=15). So it is entirely possible that only one post got unlocked in 2021 that had a historical lock before. The problem with the query is that most posts in the results get re-locked after being unlocked. I didn't check all of them but I guess it is not more than a couple of dozen that are currently still unlocked, or have a different lock that would make following useful (because some types of interactions are allowed).

@Marijn (sigh of relief) just around 12 posts?! I'll sleep easier tonight...

@SonictheAnonymousHedgehog Thanks for letting me know...