20

Preface

I support the aims of the moderation strike, but I wonder if I might add a preface as to why I haven't joined it (at least yet). For my sins, I'm the third top post editor. Curation seems now to be in my blood, and I wonder if breaking the habit would be a struggle. Also, while we the community undoubtedly see value in mutual curation, I am not certain withdrawing this labour is something the company is entirely bothered about.

So, as a curator I could cease all work and let the site's content go a little more to rot, but I worry that this action would not be the bargaining chip we Meta-dwellers would like it to be.

Finding edits

I have lots of bookmarks to find edit streams—sometimes I search for chatty phrases, which usually opens up a lot of clean-up opportunities. Sometimes I notice that a user has a long history of posts that need spelling correction or succinctification, so I also include some gentle "serial editing" in my clean-up style.

(For those interested in the community's views about editing by user: opinions are generally in favour here, mixed here, and rather furious here. There's probably a few other relevant Meta questions too.)

Main question

I am encountering a few users who are presently rolling back edits reflexively. I'd estimate they represent less than 1% of my editing work, possibly for the unfortunate reason that the bulk of readers don't care one way or the other for the quality of their posts.

I should start off on a positive note: I opened a chat with one user who was rolling back absolutely every single edit, and explained that to the best of my knowledge, my edits are based on community rules and Meta culture, and were not merely "my" rules being played out on their material. To their great credit, they agreed to fully give way, without any need for moderator involvement.

One high rep user (100K+) appears to roll back nearly every edit, but perplexingly will do so only once per post. So I gamely roll back again, and that post will then be left alone. So this one can be handled solo too.

My present challenge is a 40k contributor who doesn't seem to be keen on any community editing at all. They revert most edits, even when there are spelling errors in the text or (yikes) keyword spellings in the code. "It's my style" is the excuse I've heard far too often. I have set out the community position in a chat, and they have replied, but only to hand-wave away my patient explanations. I think readers find their technical content useful, but the grammar is sloppy and the rhetorical presentation is rambling, and unfortunately they are a prolific author. (Now I have made some suggestions on how they can improve, it looks like they are making more of the same spelling errors and awkward writing constructs just to exercise the liberty to do so.)

(A bonus theme occurs to me: are contributors who engage in edit wars taking advantage of the Wild West at the moment, or would they have been disputatious anyway? I guess this is unanswerable!)

Temporary moderation approaches

This post isn't really about one contributor, but how we can limit the damage from wilful users who are presently free to dig their heels in with no consequence or sanction. I assume all other curators are like me - six weeks of pending flags, with no end in sight to the dispute. (I have very few pending custom flags, as I think they are mostly pointless to raise at the moment.)

Normally I would put my case to a moderator, and either the user is privately requested to be willing to accept some edits without demur, or, more often, my edits are favoured with a mod rollback. In the latter case, the posts are usually temporarily locked too. Naturally, I am always open to compromise and negotiation, but without mod backup, tempestuous contributors don't approach any site dispute in the same way. (They will just do what they want, and they do not usually deliver their opinion to me with any sort of diplomacy.)

I will answer my question with a couple of starter ideas:

  • Open a chatroom with the writer and put the "mutual editing" case forward with civility. Users who reflexively rollback do tend to be disputatious anyway, but as above, this can produce surprising results.
  • Could editors bandy together and rollback any rollbacks where they see an edit they agree with? Maybe a chatroom or other means of collaboration?

It may be that we have limited options without moderation, and that we're just discovering that the above is what happens when you don't have your "human exception handlers" at the ready.

I appreciate that my enquiry here may be answered with a "let the site burn and join the strike", and if that is your view, I'm willing to hear it.

21
  • 40
    "I am not certain withdrawing this labour is something the company is entirely bothered about" -- that's kind of the point tho. There are two possible outcomes -- either they wake up, or we go do this free work for someone who actually appreciates it. (And from my perspective, it's looking more and more like the second options is the wise choice)
    – Dan Mašek
    Commented Jul 24, 2023 at 23:10
  • 45
    Flag it as usual. There's nothing else to be done. If it escalates to code of conduct violations, use the "contact us" form. Otherwise, it'll be handled after the strike, possibly sooner if it attracts more flags Commented Jul 24, 2023 at 23:21
  • 9
    Given the massive slowing/ stoppage of moderator-required moderation actions from the strike, whether someone is part of it or not, there aren't currently a lot of options for responding to others abusing the system other than flagging for mod attention and sitting tight. I don't believe mod flags ever expire, so it will get handled... eventually. I wish there was more to offer, but that's also kind of the whole point, I suppose.
    – zcoop98
    Commented Jul 24, 2023 at 23:26
  • 9
    @Dan: the company harming the site/ecosystem in order to prioritise the monetisation of LLM building is my great worry. But I don't see there are any viable alternatives for decent contributors and curators to go to - either the company doesn't understand the complexity of the Q&A model (Yahoo Answers) or they aren't even suited (Reddit) or they aren't adequately funded (small community efforts from the last company/community bun-fight).
    – halfer
    Commented Jul 24, 2023 at 23:42
  • 3
    (I assume the downvotes here are from folks who don't like my view that the withdrawal of non-diamond curation isn't a great bargaining chip. I'd like to be wrong about that of course, but more importantly, that's not the point of this post).
    – halfer
    Commented Jul 24, 2023 at 23:43
  • 3
    We're trying build a library of high-quality questions and answers. New members may not realise that yet, but long-term contributors should. IMHO, if someone persists in posting sloppy answers, and actively fights attempts to improve the quality of their posts, then we should mark those answers as "not useful" by downvoting.
    – PM 2Ring
    Commented Jul 25, 2023 at 9:59
  • 1
    supposedly not everyone is on strike so just raise your flag and wait for someone interested to look at it
    – user13267
    Commented Jul 25, 2023 at 10:10
  • 14
    Fun fact: Stack Overflow is not a forum. Commented Jul 25, 2023 at 16:20
  • 2
    Like all strikes, the core idea behind this one is simple: Let the trolls wreak havoc in order to gain visibility, so that the owners begin to act responsibly. As such, I agree with the main answer of just behave as normal and let the site blow up as the mods strike. That's the idea since day 1 of the strike.
    – Alejandro
    Commented Jul 25, 2023 at 17:43
  • 3
    @BenWatson: it's not pedantry - you just missed a reference to a common Stack Overflow meme. See this search. It is very common for denizens here to say that Stack Overflow is not a forum, since it is a Q&A site, and is intended to work differently to a forum. The site "not being a forum" is in the official Help documentation.
    – halfer
    Commented Jul 26, 2023 at 15:38
  • 3
    @BenWatson: this answer is perhaps the best explanation of why mutual editing exists, particularly the bit about Wikipedia. The community is looking to curate pages rather like documentation, rather than to host a chat/discussion.
    – halfer
    Commented Jul 26, 2023 at 15:51
  • 1
    Yah. that pretty much covers it. Considering the other curation alternative is to leave a comment to ask the author to fix the post and downvoting it when they don't, allowing third party edits is pretty friendly. And of course there is nthe non-curation alternative of do noting, but that doesn't help the site in the long run. Commented Jul 26, 2023 at 17:15
  • 1
    I can, but it depends on what Wikipedia considers enough of a "forum"; the definition given at the start of the page is the following: "An Internet forum, or message board, is an online discussion site where people can hold conversations in the form of posted messages." This does not describe Stack Overflow of course, but opinions there may differ; the interpretation actually used seems to be pretty broad. Commented Jul 26, 2023 at 17:30
  • 2
    I'd agree with @Darren here - the Stack Overflow definition should hold on Stack Overflow, but people outside of the SO community will have a broader definition. Ben links to the article in good faith, but it doesn't apply to us - the author is not likely to be claiming community expertise. I think it is fine to leave it be.
    – halfer
    Commented Jul 26, 2023 at 17:41
  • 1
    Good points. A reversion war on someone else's site is not how I want to spend my Wednesday. Commented Jul 26, 2023 at 19:20

2 Answers 2

61

Flag objectionable content as usual

...and don't worry if the flags aren't handled.

It's clear that the company either doesn't care about the extent to which Stack Overflow succeeds in building a library of detailed, high-quality answers to every question about programming - or if they do, their concept of "high quality" is beyond inscrutable to mere mortals like us, who value simple things like "direct answers without noise" and "titles that accurately and specifically describe the issue" and "clearly identified canonicals that actually get routinely used to close duplicates" and "easily understood, proper MCVEs"... I could go on.

On the available evidence, they won't be bothered if questions and answers degenerate into a discussion, LLM-generated text features everywhere, nobody ever makes appropriate attempts at debugging before asking, etc.

After all, their incentives are clearly not aligned with any of that, and their KPIs don't have a hope of capturing it. People saying mean things about Stack Overflow off-site doesn't really impact that much on ad revenue. (Besides, they largely say mean things because of policies that we'd like to do more of, and think we aren't enforcing nearly well enough.) Closing questions means they don't get fresh answers and they might even get deleted, reducing the volume of pages that the site could be using to pull in more search traffic.

It also means that when people search for beginner questions and get swamped with a bunch of nonsense from the Internet being the garbage heap that it is, they aren't seeing a whole bunch of stackoverflow dot com links in those search results. Actually resolving the issue? That doesn't pull in money from advertisers. Page views (thus, ad impressions) do, and Stack Exchange Inc. has a vested interest in building on that total (something like 70 billion just for question pages last I checked; apparently the SQL implementation on data.stackexchange.com uses 32-bit types for SUM results, so it takes a little finagling to determine).

However, the company does seem to care quite a bit about the code of conduct, considering the effort expended to revise it recently without a proper community process, even though they know how poorly that went for them in 2019. And honestly it's not even that bad as policy goes; nobody really wants rude or bigoted - or even just off-topic and rambling - content on the site. And certainly there's a clear business case that getting rid of rude content encourages more participation, which is good for business.

And we've already seen, through multiple recent incidents, that staff have been handling the flag queue. Currently they're being expected to pick up the slack for the striking moderators, which has got to be awfully burdensome. I can't imagine there are very many staff members available for a job that already strained a couple dozen volunteers.

If we can't convince Stack Exchange Inc. of the value of human moderation (and the non-value of LLM-generated content) by showing them what the site becomes otherwise (because they apparently don't see an issue with the result), we should try to convince them by showing them how much valued work moderators do (and how much of an additional burden is created by allowing miscreants access to that LLM-generated content). The length of the flag queue is the most obvious metric for that.

Don't waste time on what can't be seen

Lots of curator types are on strike, too. Close votes can't be relied upon, except via Mjollnir. Downvoting and dupe-hammering stuff takes eyes off it, and that change theoretically reflects in view counts etc. Simply voting to close won't accomplish a lot in most cases. Better to let sub-3k users flag things - that way it also generates a flag.

Deliberately getting into an argument won't reflect well on you, so I can't recommend it. So it would be better not to waste time on edits at the moment. Even if the company doesn't care about you withdrawing the labour, at least it's labour you're not doing, voluntarily, for a company that you presumably see as also not caring about you. (From the perspective of strike pressure, the best case outcome is getting into an argument incidentally and thus having something to flag. That's, well, still ugly.)

Time not spent on improving or contributing site content, is time that could be spent on publicizing the issue on social media, writing about it on a personal blog, or contributing to alternatives that will take a firmer stance against LLM content and enable their moderators and curators (or equivalents) to remove it aggressively. If Stack Overflow becomes the place where AI hallucinations are commonplace, while other sites are the home of freshly written expert advice, eventually that will be reflected in user demand.

12
  • 27
    There is an upside to the staff moderating the site. They might fix/improve the moderation tools. Nothing improves the "taste" of software quite like a regular diet of dogfood. Commented Jul 25, 2023 at 0:49
  • 6
    @user4581301 - What is more likely is they will continue to make amateur moderator mistakes. As more time goes by, and more and more LLM content is unofficially officially allowed on this network, the more apparent it is that perhaps it’s the entire network that is the problem. Unfortunately, there isn’t an actual alternative, but that fact won’t keep answer authors like me around forever Commented Jul 25, 2023 at 9:28
  • 9
    @SecurityHound Not of this scale but Codidact does exist. Quite some Stack Overflow regulars can already be spotted in the wild there.
    – Gimby
    Commented Jul 25, 2023 at 11:22
  • @Gimby - I have seen the question’s being ask at that community, at least at the community that would interest me, and they are ridiculously trivial and wouldn’t provide a Intellectual challenge Commented Jul 25, 2023 at 11:53
  • 1
    @SecurityHound you find intellectually challenging questions on Stack Overflow still? I have to wade through dupes and "I am throwing things at the wall and see what sticks because reading documentation is boring" questions. Maybe if I do that for an hour I find something that is worth the effort.
    – Gimby
    Commented Jul 25, 2023 at 12:40
  • @Gimby - I wasn’t talking about SO specifically, SU, and Power Users. SO became a toxic fire dumpster awhile ago in my eyes. So I don’t disagree that SO can’t provide the intellectual challenge either. Commented Jul 25, 2023 at 12:49
  • 12
    I don't know about you guys, but I don't care at all if the questions are "intellectually challenging". I would go to codegolf.se if I did. Rather, I want questions on a SO replacement candidate to be high-quality, suitably-scoped prompts for high-quality expositions of crucial concepts. That includes trivial ones. If, say, Codidact can comprehensively explain how relative imports work in Python, correctly identifying mostly-unnecessary hacks as such, without dozens of near-duplicate questions that each have dozens of overlapping, mostly outdated or wrong answers, that's a massive win. Commented Jul 25, 2023 at 19:45
  • 10
    Seriously, I would rather have a few thousand questions of that caliber, than the few million Python questions that Stack Overflow actually has. If anything, the most important concepts for a given programming language or technology are mostly not that "intellectually challenging" - having to put up with "challenges" drastically reduces the mass appeal of those languages and technologies. Commented Jul 25, 2023 at 19:53
  • @KarlKnechtel The idea is enticing indeed, just deal with simpler things. The whole dupe closing process on Stack Overflow is tiresome and demotivating. Just let me answer stuff in my way and be damned that an answer which words it slightly better may or may not exist somewhere else. People don't search anyway, so who are we making the effort for. but maybe the underlying compounding issue is that the whole state of the site makes a person pretty jilted.
    – Gimby
    Commented Jul 26, 2023 at 12:09
  • Thanks Karl, this is good food for thought. I agree that it is not worth getting into an argument deliberately, though I would posit that where a curator approaches a troublesome contributor for a discussion, an argument will only happen if they choose to steer things that way. A conversation that features amenability and compromise ought to be the norm, and my foolish optimism wants to assume that anyone I approach has the capacity to be reasonable.
    – halfer
    Commented Jul 26, 2023 at 17:29
  • 1
    One thing that we should all have noticed is that those trivial questions, when asked early enough to not be duplicates, are the ones that rack up the most upvotes. Why? Because they help the most people. That super-challenging esoteric question gets a great rush of instant votes from the folks who read and appreciate the problem, but after that it's the folks who look for it, or something similar, who find it and and upvote it, and face it, there aren't that many lookers. There needs to be room for both, but there needs to be rigorous curation to remove the noise of the trivial duplicates. Commented Jul 26, 2023 at 18:43
  • Or a system that can eliminate the duplicates from being asked in the first place. AI looks to be the start of such a solution. Commented Jul 26, 2023 at 18:45
16

If anything, your concern is another great reason to be less liberal in your curation activities.

Simply put,

It is not safe.

Even when you take your voting and editing privileges for granted, moderators always played the role of exception handlers when some unreasonable user started pushing buttons the other way. And sure the advice to flag and disengage is great. But sometimes the other end won't stop, leading to a stream of abuse. And the only way to protect yourself from this is to stick to anonymous forms of curation, such as flagging.


Regarding the strike, know that it's a rollercoaster of emotions for me to leave so many inappropriate posts and comments unattended. I've seen name calling, insults, even denigrations and demonizations towards specific users! But let's face it, these things are what happen with insufficient moderation. With each bystander calling it out, whether it be via the Contact Us form, or on social media, it ought to send a message to Stack Exchange Inc that the company can't handle this without the community.

9
  • 6
    I have a general distaste for the broad conception of "safety", and the associated theory, that underlies this answer; but otherwise I absolutely agree. Commented Jul 25, 2023 at 9:36
  • 5
    I honestly do not see where that distaste is coming from, @KarlKnechtel. The code of conduct explicitly outlines the expectation of fostering a "safe and welcoming community". Being safe includes having exception handlers ready to handle violations quickly.
    – E_net4
    Commented Jul 25, 2023 at 10:15
  • 11
    "know that it's a rollercoaster of emotions for me to leave so many inappropriate posts and comments unattended" - IMO Stack Overflow 2023 does not deserve that anymore, such emotions are a relic from times where the site was something worth making sacrifices for. The way that it is now deserves a healthy amount of indifference, for your own mental stability. It is not your baby anymore.
    – Gimby
    Commented Jul 25, 2023 at 11:29
  • 2
    I'll surely work on that, @Gimby.
    – E_net4
    Commented Jul 25, 2023 at 11:39
  • 2
    I wish you good luck. I'm a nostalgia hound myself, I like to pretend it's still the 80's and 90's. Letting go of the past is not easy.
    – Gimby
    Commented Jul 25, 2023 at 12:42
  • 2
    (The distaste comes from broader personal, political views that are not particularly relevant here - I just don't like letting the matter pass silently. As you may guess, I don't like how the code of conduct phrases the point, either; but I accept that it's commonplace and that you are using its framework.) Commented Jul 25, 2023 at 19:50
  • I appreciate this answer, thank you. Just a quick remark on the interesting meta-meta topic of safety; I think this might specifically be psychological safety, which is in fact not new - the term apparently was coined in the 1950s. I note that the Western world is in a time of particular cultural flux, and that when people learn new things, they necessarily attribute them to the flux, rather than learning that they didn't know a thing, and adjust their world-view accordingly.
    – halfer
    Commented Aug 1, 2023 at 17:54
  • I don't aim that remark at Karl as such - this is more an idle commentary on how we all might hanker for a mythical past, when ideas were simpler, or there was less confusing choices, or there was more agreement about things, etc. Of course, every generation makes the same cautious judgements about "new" ideas, but I suspect the halcyon era is an illusion.
    – halfer
    Commented Aug 1, 2023 at 17:58
  • On the nostalgia of Stack Overflow or what it was, I confess that may be a struggle for me as well. There has probably never been a community of strong bonds here as such, though the site has always been modelled on a city, not a village. But I perhaps should ponder this some more, even if means a win for the vandals and help vampires in the short term.
    – halfer
    Commented Aug 1, 2023 at 18:02

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .