I responded with an explanation as to why OptimizationVariableInterface is different, and indeed complementary, to what is implemented in Optim for manifolds. It’s a very helpful explanation, that I edited myself. It’s been hidden twice by a moderator. It’s a shame, because I think it’s really helpful information for the community.
To the @moderators
I consider it a very extreme measure to hide somebody’s message, especially if it happens in an otherwise very informative and civilized conversation. Please provide more specific information for the rest of us to understand your decision.
I think what happened is that someone flagged it and it was automatically hidden until a moderator looks at it?
(Unfortunately, these days people see a longer, structured post and immediately suspect that it is AI-generated.)
Then the system is broken. Flagged posts should not be hidden until after a moderator has checked it, or perhaps a threshold of flags is met. And those people who abuse the flagging should be sanctioned, e.g. lose the privilege. Otherwise this opens the opportunity to people hiding information to protect their own interests, which is what I suspect is going on here. And it’s very sad to see.
I think most people here are doing the best they can and working in good faith; your own post was restored after a short delay.
(Please realize that, behind the scenes, there is an unending deluge of AI-generated spam and other AI-generated posts in contravention of the rules in this forum, requiring continual moderation effort. Many people are doing their best to keep this a usable site for human discussion.)
@mbauman knows the technical details of moderation much better than my superficial knowledge, however. (I just see a daily queue of moderation requests.)
I agree. But it only takes one, doesn’t it? What about when someone isn’t? Is the system robust to that? It doesn’t look like it. Unhelpful people hide behind dismissals like this. So why not make an effort to make the system more robust to that, rather than dismiss concerns as “It’s basically not an issue”?
Plenty of solutions:
- Leaving flagged posts visible until a moderator has checked them
- Requiring several flags for a post to be hidden
- Sanctioning flaggers whose flags get rejected
- Making flaggers public to the moderators, or even the poster?
Or, as @stevengj suggests, is it simply not an issue?
As always, there is a trade-off between type I and II errors. Yes, I know it feels weird when one’s perfectly OK posts are flagged, it happened to me too, but you should not take it personally. They are usually restored.
Feel free to make that effort. If you end up designing a better system with the same amount of personal attention required, you should approach the maintainers of Discourse, so they can incorporate it into the next update. This forum is just a user of the engine so it is not the best place to discuss improvements.
I think it is quite a leap to come to that conclusion. As @stevengj said, the more likely explanation is that someone thought the post ran afoul of the guidelines on AI posting (under “Keep it Tidy”).
All of the solutions mentioned assume that the flagger is acting in bad faith, which seems very unlikely in the vast majority of cases.
Publicly revealing the users who flag posts would only serve to discourage people from flagging at all. Imagine a case where two people get into a heated argument and someone throws a flag after a “bad” post. Having that be public only serves to let that person be harassed in private DMs because now everyone knows who it was.
And leaving truly bad posts up for a longer period of time until they hit some criteria in the off-chance someone in the future tries to hide information seems like a bad trade-off.
Also, flags aren’t permanent, so if one is placed and the moderators disagree they can just remove it and remedy the situation. But in the end moderation is always going to be subjective no matter how “robust” the system is.
The people here dismissing my concerns have no context whatsoever, whereas I do. Please recognize that. In my view these comments go too far.
All other points gratefully received. Thank you for understanding.
Here’s the way the system works, for better or worse. Anyone can flag any post for any reason. A single flag on its own does not automatically hide posts — usually! Posts are only hidden if a moderator agrees with the flag and chooses to hide it OR if the post has accrued multiple flags without any moderator action. The number of flags required for automatic action before mod review varies and is quite opaque, even to me… but it depends upon the flaggers’ trust levels and number of times mods have agreed with their flags in the past and the flag itself and maybe more. We have our site settings set to the default “medium” hide post sensitivity which is what tunes this.
Now once a post has been hidden, the original poster gets an automated message urging them to edit it. Any edit will restore their post. Mods aren’t notified of this and we don’t get notifications to review or approve edits. But now any single re-flag of the post by anyone of any trust level will cause it to be automatedly hidden until a mod reviews the flag. This is kinda-sorta described in the automated direct message you received.
Because this post has been hidden more than once, your post will now remain hidden until it is handled by a staff member.
We can see how it works. But you say there are settings, so we can change this, for the better…
Indeed, the context of replies is typically limited to what is available in the threads themselves. So, if you have any additional context that someone is intentionally messing with your posts, then that seems like something you should bring up privately in DMs with the moderators.
Not clear what the moderators can see or do about this. Other than moderate…
Moderators have tools to temporarily (or permanently) restrict people’s ability to participate in the forum altogether, depending on the nature of the offense. I haven’t seen it happen often, but it has happened.
Whether or not that happens in this case would depend on the circumstances and evidence, obviously.
I don’t think the behaviour I’m talking about really meets that bar. It’s easy to say “yeah, I thought it was spammy”. So it isn’t a solution.
Oliver, I can say quite definitely that you are not being targeted for flag harassment. In the past three years you’ve had exactly one post flagged: this one. It was flagged by one person as AI and I agreed. You obviously manually edited it which unhid it (and I saw and appreciated it!), but it still retained some claudish which prompted another (different!) person to flag it again. This caused the auto-hide pending review. And upon review, Steven disagreed with the flag and un-hid it… even though I suspect he may not have dug into all the context of that first flag and edit.
Yes, the system is tuned to be very sensitive to reflags but I’d say it’s for very good reason! Lots of times inappropriate (or worse) behavior is un-hidden by a bad-faith no-op edit. It’s quite rare for this particular trip-wire to be a false positive.
I do think it’d be good to detangle the spam and AI flags. I’d appreciate feedback on a change to the site I proposed a month or two ago:
I don’t think the number of times my posts have been flagged in the past is a strong signal as to whether this post was flagged out of malice; also you are missing other context. However, I can accept I have been caught by the reduced requirements of the second flagging, and likely just a random, rather aggressive flagger. Good to know. Thank you.
I’ve commented.
I only see what I can see. Larger conflicts and problematic behaviors spanning multiple venues throughout the community at large is precisely the reason-for-being for the Stewards.