The original post uses “roll-up” instead of “catch-all” for some reason.
There is a long-festering problem in some tags where some questions are closed by dupehammers, using a single roll-up question as the duplicate target. A “roll-up” question is defined here as a question trying to cover multiple minor topics within one question and a set of answers. So this Java question about null pointer exceptions does not qualify, as it is about a single topic.
A prime example would be this regex roll-up which has a large number of duplicates. This was by design.
Questions that are clear duplicates, but you can’t find one quickly.
To be fair, PHP and other tags have such roll-ups (example), and I have participated in hammering them as such. And there are a lot of questions that are low quality, where the temptation is to simply close them as the duplicates of the roll-up. I mean, it answers the question, doesn’t it?
The problem is that this has started to promote two undesirable community actions:
Lazy closure
Dupehammers are a “one and done” action. Moreover, there is a belief is that these questions answer all the “core” elements and are therefore “useful” in low quality situations. The question for regex theoretically covers all symbols used within, so why isn’t that useful? But this type of closure assumes that the roll-up covers all cases. The danger of dupehammers has always been that the target question doesn’t really cover a specific use case. Lazy closure doesn’t even bother to find that out first. Thus it becomes the action of choice for dupehammer users. It’s problematic, but the community largely self-regulates this so it’s not been a major issue. A low quality question can be closed for many other reasons beyond duplicate.
Tag gatekeeping
This action is the more problematic one. What we’ve been seeing for some time are “brigades” (for lack of any better term) of users who are committed to ensuring that only questions they see fit in a tag are open. Thus we get a number of these:
What this has turned into is not laziness, but deliberate actions, where we see the same users doing this over and over. Or, to quote a comment under the question I got the screenshot from:
I invite readers to examine the earlier question and ask themselves if any question could possibly be a duplicate of that question. If the answer is “no”, please vote to reopen (and leave a comment giving your reasons for doing so). Closing this question, in this way, is sending a clear message to Peter, the OP (the polite version): “get lost”. This catch-all closing of questions having a “regex” tag must stop.
I don’t know that it sends a “get lost” message, as much as it sends another message moderators have been fighting against for years: RTFM. What these roll-ups have become, in essence, is another “fine” manual for users to read. Duplicate closure like this is basically throwing a volume of information at users and telling them “Figure out what, in this giant pile of information, answers your question.” That’s not useful.
It also effectively acts as a veto for anything any dupehammer user sees fit to close it as. Roll-up questions worked well as a philosophy for a long time, but (as the old saying goes), this is why we can’t have nice things.
The rule
The rule would be as follows:
Roll-up questions are useful in general, but may not provide enough guidance to users with specific questions, and serve as poor signposts to users looking for specific answers. Please use only specific questions for duplicate closure.
FAQ
- Moderators would enforce this new rule. No system changes would be made.
- Moderators would find out about violations via flags. Moderators already get an autoflag for closure disputes, and users could flag instances of this rule being violated.
- Enforcement would follow standard enforcement: A warning on the first offense and suspension for subsequent violations.
- Any other duplicate closure would still be allowed. If someone feels strongly enough that it’s a duplicate, they should go find that specific question. Moderators will still not solve duplicate disputes, but the list of roll-up questions isn’t long, and it’s a fairly objective standard to enforce.
That’s actually a good theory. Given that LLMs, even if not always correct, are at least polite, they are a serious competitor for these “competitive” sites like stack overflow.
The problem is that those models are trained on the data from stackoverflow. Without humans to ask and share modern solutions, the answers from the models will stagnate and become outdated.
Maybe the LLM trainers will then invest some energy in creating a stack overflow alternative for people with basic social skills. Or well, that’s maybe too harsh, but I feel like stack overflow incentivises antisocial behaviour.
They would be better off if they would also allow “conversational” problem solving. That means anyone can ask questions, without being forced to ask a “high-quality” question. Missing details or similar questions can then be discussed in the conversation, instead of just closing the question.
Well, now I am reinventing web forums.
I think you’re succumbing to the belief that if a solution isn’t perfect then it doesn’t work. That’s not how things work. You can have incremental improvements that replace a big problem with a smaller problem, and a smaller problem is easier to solve.
Also, StackOverflow already suffers from old and stale replies for years, and that’s not hurting the site. Oddly enough, that’s a problem that’s mitigated with the way that data is queried,and that’s handled quite well with large language models.