while making it easier to edit and reopen closed questions
Please keep in mind the need in this case to make it easier to re-close questions, too. More on that below.
Today on Stack Overflow, roughly 20% of questions are edited after they are closed and just 3% of closed questions are ever reopened. We’d like to see more questions improved upon so that they have a better chance of being reopened and answered.
Has there been a study done to see what of the 97% of un-reopened questions (or 80% of un-edited questions) you think deserve to be reopened? E.g. is there an actual problem with only 3% of questions being re-opened? It could well be that that's actually a good number of questions to be reopened from a quality perspective.
Askers are often unsure how to get their question reopened and onsite guidance can be confusing. They often experience a pile-on effect of downvotes and comments while trying to edit their question in parallel. This can feel overwhelming and intimidating for question askers and drive-by visitors.
Some better guidance and/or handholding for askers whose questions get closed (by others; self-accepting a recommended duplicate should be left out of this basket, as one example) would go a long way. The other hurdle is getting askers to read the guidance already given to them...
Curators are burdened with an enormous backlog of tasks. We want to reduce the burden by helping question askers improve the quality of their question through product guidance, relieving curators of this tedious and often thankless task.
Part of the reason we have an enormous backlog of tasks is that we are artificially rate-limited at many curation tasks. Specifically, the number of close votes we can cast per day and the number of review items we can cover per day. If we were given 100 close votes per day instead of 50, or if the Close Vote Queue's pile of 40 close votes were separate from our pile of 50 close votes per day, I imagine the backlog would drastically decrease.
Shog long held that the reason for never* upping these limits was to prevent burnout, but I have never bought that. Additionally, these burnout concerns somehow don't matter when it comes to moderators, who are allowed to do as many of each task as they want.
I hear some people also routinely run out of flags. While I do use flags almost every day, I personally have never gotten close to running out once I hit 100 flags per day. I'm guessing those that do are probably members of Charcoal or power users of SmokeDetector, at least.
* There was one test of raising the limits a year or two ago. As far as I remember, the test was a pretty clear success. Unfortunately, an increased cap never got implemented in a more permanent manner.
We will also be testing automatic reopening
When a user edits a hidden question in a substantial way, it will automatically reopen (unhide) the question and return to its pre-close, public state. Additionally, a question can only be automatically reopened once.
Oh god no, please don't do this. There's no good reason to open a question automatically. If you do this, you have to let those close voters who closed the question vote a second time, and you should notify them about the question being reopened, too.
putting the onus on the question asker to improve and reopen their question and explore the notion of “hidden” as opposed to “closed” questions, giving askers the space to (somewhat more) privately improve their questions. More details on these concepts are below.
We’d like to reconceive “closed” as “hidden” so that users can improve their question without feeling embarrassed or exposed. It’s a gentler way to remove a low quality question from general visibility while giving the author a chance to edit it. Privileged users (mods and high rep users) will have a path to access these questions to provide additional coaching and guidance.
I think this is a pointless change made under the (failed) guise of "being nice", personally. Telling me my question is hidden isn't any nicer than telling me my question is closed. However, I think it is a good idea to hide closed questions from question pages/lists for users under at least 3,000 reputation.
To maintain quality, if hidden posts aren’t edited within a certain timeframe, they will be deleted. And to mitigate abuse, we’d like to explore dynamic vote thresholds for closing based on age so that newer questions require fewer votes to close than older questions.
Can you detail these criteria, specifically how they compare and contrast to the existing conditions for deletion (e.g. the Roomba)?
Phase 3
- Score and voting capabilities are removed.
Please don't remove the ability to vote on a question while it is closed. If the reason you want to do this is to prevent an overwhelming number of downvotes, simply implement a floor for the displayed score... e.g. -1 or -2 or something like what was tested last year by Shog (successfully, I might add). Downvoting is one of the most important curation tools that we have. It also affects the ability to delete vote a question or whether the question will Roomba.
#Some things that weren't covered by this question:
- Please consider adding more regex checks (yes, yes, I know regex is dangerous and bad all you programmers reading this) for questions at the asking phase for off-topic reasons. Some examples of phrases in a question body or title that should at least show a warning banner (saying something like "opinion-based questions are off-topic. please do not ask them") before a user is able to ask the question:
- best practice(s)
- idiomatic
- how should I
- what do you recommend
- what's the best
etc.
Probably somewhere in the region of at least 40% of the 40,000 close votes I've cast on Stack Overflow were on questions that contained one of these phrases or similar.
- Is there any effort still on expanding the close reasons for which gold tag badge holders are able to unilaterally close (Mjolnir) a question? It's been wanted for years, and we were just starting to get a hint that it might be coming last year as Shog started testing changes to the closure process.
