[X] My Assistant

I don't think the existing wording mentions deliberately incomplete galleries but that should be easy to add explicitly.

Not for the definitions listed in the OP, yeah, but we are to:

I'm fine switching the wording from "downsampled" with say "degraded", with some examples listed, including "downsampled", "blurred", "pixelated", "left out significant core content" etc.

...then those newer spams will be covered.

It also strikes me that some of what is covered in this new role is partially covered by FC so that might also be worth re-writing and rolling those here (intentional censorship for example)

These two will probably be removed from "forbidden" if we roll with option 2:

- Intentionally adding censorship beyond the artist's.
- Having tiny or barely visible images.

in theory this issue exists with current petitions as well as any gallery can be expunged for any made up reason if one can get enough votes

Key difference is that the other categories are objective, this one isn't.

I think a crux of this issue is that currently, for MP accrual, expunging has practically zero direct contribution. (with the exception of awards) All the top expungers got their MP by completely unrelated activities to expunging some even partially paying for the privilage, you can get to 20MP purely via donations, hentai verse and keeping an inactive account for ~3 years, with enough seed minutes you could even get to MP 29 in that time simply via the EH Tracker again not even necessitating reading the rules, let alone applying them.

I would like to see expunging and renaming play a role in MP, but that's probably beyond the scope of this discussion.

CODE

"Content has been defaced in a way that significantly reduces the value to the community, by either the addition or significant negative alteration of core content. The addition of Non-Content in a way that is not currently exempt and/or is excessive is also forbidden."

This would make an already subjective matter even more subjective.

E: This is an extreme example mind and I don't exactly have a suggestion for what might be better other than a whitelist and I kinda agree that might be somewhat tedious though I would argue that the new whitelist tags offer sufficient precedent that this approch could work, and since you already have the comparison and expung vig threads to go for you probably have enough information to go for when deciding if someone with say 15 mp should be allowed to make subjective expunges.

I dislike the idea of making a whitelist every time something like this happens. Before there was one for rough translation, now there could be one for defaced petitions, what there will be next?

Not for the definitions listed in the OP, yeah, but we are to:...then those newer spams will be covered.

These two will probably be removed from "forbidden" if we roll with option 2:

That is probably good enough I missed that post in the flood.

Key difference is that the other categories are objective, this one isn't.

Not sure if that matters for detering sockpuppet accounts is all I am saying, after all the concern is they would start petitions eroneously in the first place and vote for them with their own accounts. The main reason this doesn't happen than much currently is because it is possible to ban such accounts faster than they can get to the required MP to do this more than once. There is basically a diminishing return in MP requirements in terms of deterrence for certain kind of abuse mechanically.

This would make an already subjective matter even more subjective.

The idea is the examples should sufficiently clarify the scope of the rules in the same way as other expunge reasons work. I don't think it's so much of a stretch personally. There is of course a delicate balance behind making rules robust enough to stand up to abuse and making them clear and easy to interpret. Subjectivity is not inherently bad in such cases as long as there is some common standards that are applied. But that in itself is an entirely subjective matter and one that is entirely up to the moderation to decide.

I dislike the idea of making a whitelist every time something like this happens. Before there was one for rough translation, now there could be one for defaced petitions, what there will be next?

I don't think creating a whitelist for subjective matters that require either a) Domain knowledge (such as the one required to judge a translation) or b ) Weigh in on a subjective matter that requires understanding of nuance and context is really that problematic. The precedent for this is already set, and I don't think there is that much danger in that this scope relatively narrow anyway.

This post has been edited by Theonik: Apr 22 2022, 05:51

Not sure if that matters for detering sockpuppet accounts is all I am saying, after all the concern is they would start petitions eroneously in the first place and vote for them with their own accounts. The main reason this doesn't happen than much currently is because it is possible to ban such accounts faster than they can get to the required MP to do this more than once. There is basically a diminishing return in MP requirements in terms of deterrence for certain kind of abuse mechanically.

If people are angry when dealing with an expunge petition, they are significantly more likely to game the system. That combined with the certainty of a tracker flood is why I said a mechanical solution is needed. If people want a lower initial MP requirement, then this rule cannot be applied retroactively, because we would get too overwhelmed.

Bottom line is: We could consider a lower MP requirement later. But at its initial stage, a higher value is required to make things more manageable.

The precedent for this is already set, and I don't think there is that much danger in that this scope relatively narrow anyway.

That tag is just a very special case, not really meant as a precedent. We kinda don't want a special whitelist for every single abusable thing out there. That can get rather tedious to manage eventually.

This post has been edited by PrincessKaguya: Apr 22 2022, 06:16

this scope relatively narrow anyway.

Fairly limited in terms of uploaders, stupendously large in scope of galleries if this is to be applied retroactively.

That is a big part of the issue here, lower MP requirement at the start would be very beneficial for catching up on that backlog, but also indeed the riskiest moment to do so because it'll be awfully easy to slip a malicious expunge through the cracks when this shit is going to flood the tracker.

I'm curious, because the "backlog" would be such a massive undertaking to catch up to if this is indeed intended to get expunged as well, would it be possible to have a temporary whitelist for trusted users below the MP requirement so that they can still help spread the workload out more? similar to how being on the Rough Translation whitelist lowers the MP requirement for Rough Grammar from 15 to only the 10 required for whitelist?

This post has been edited by Glovelove.: Apr 22 2022, 06:18

I'm curious, because the "backlog" would be such a massive undertaking to catch up to if this is indeed intended to get expunged as well, would it be possible to have a temporary whitelist for trusted users below the MP requirement?

Folks can always post a request over the expunge thread, so I kinda don't think a temporary whitelist is really needed.

Given the initial cleanup will probably mostly focus on the uploaders that triggered this discussion I'm not so worried about people being unable to find the galleries without me dropping them off in the thread, more so getting burned out after the 2000th expunge in a row knowing they're not even halfway there yet.

This post has been edited by Glovelove.: Apr 22 2022, 06:43

In terms of the backlog I wonder if considering how few uploaders make up the bulk of the current backlog of several thousands of galleries is it not an option for admin to do a one off mass expunge delete without requiring a petition? I don't see any way 2000 petitions going through the expunge system is going to not overwhelm either the expunge thread, the logs or people's patience to get through this.

It's awkward if not impossible to automate. Can't do it by the scanmark tag, can't do it by resolution, can't do it by rating. Even for cgc's galleries, not all may be bad enough to warrant the expunge. A lot of his older ones weren't that terrible.

Initial tracker flood might be an issue, but I would rather that than permitting those older "dd dolls" BS to stay.

Most active expungers have 20+ MP. For this subjective expunge category, we probably don't want anyone but the experienced to touch it, so I would say 20 is the optimal number.

(Tbh, even 25 MP could still work just fine, since the expunge thread is always a thing if people don't have enough power)

We can also start this gig with a higher MP (i.e. 25) initially, just to test the water. If misuse has proved to be not a frequent occurrence, then we can drop it to 20 then 15.

I agree with this. At the start we should be looking at 25 while dealing with all the onslaught. After it becomes stable we can look at 20 (or even 15).

I don't think I've ever seen an expunge pass without someone 25+ voting for it. And nearly all of the regular expungers are 25+

It's awkward if not impossible to automate. Can't do it by the scanmark tag, can't do it by resolution, can't do it by rating. Even for cgc's galleries, not all may be bad enough to warrant the expunge. A lot of his older ones weren't that terrible.

But you may be able to just say the last 1 year gets expunged and you revert the few (if any) galleries that aren't in that category or a combination of age and tags. It may be worth considering if it cuts the backlog by a few thousand with a few appeals coming out of that.

If it's a multiple of 1000, then any wrong ones are very unlikely to get reverted in a reasonable time frame, if ever.

omg, i can't express how happy I am seeing this topic

It is real pain after reading all the guidelines and knowing it is currently not okay to report people for scamming and adding ads onto works THAT THEY TOOK FROM EXISTING GALLERY.

I'd agree that the former code seems to be on the safer side, since "intentionally downsampled" seems like it's potential of being misused is real high

And many translator tends to downsample the images not with malicious intention, but mainly because their available internet simply does not allow them to upload them in full size without constantly crashing.

thanks for considering this new expunge reason. for real.

If it's a multiple of 1000, then any wrong ones are very unlikely to get reverted in a reasonable time frame, if ever.

Well yes, that's exactly why I agree high MP requirement isn't a bad thing initially but also worry about the more exclusive club of high power vigs getting overworked so we should try to find a solution that can address both potential issues instead of calling it close enough when we have sort of a solution for just 1/2

I'd agree that the former code seems to be on the safer side, since "intentionally downsampled" seems like it's potential of being misused is real high

And many translator tends to downsample the images not with malicious intention, but mainly because their available internet simply does not allow them to upload them in full size without constantly crashing.

Didn't know this was ever a reason for downsamples, honestly sounds a bit pathetic to permanently have something at lower resolution because of an obstacle that should be preventable (like emailing it to a friend with better internet connection or the like) but I can't imagine it being a more common cause than uploaders not knowing/caring how to get full resolution raws.

But yes these shouldn't be expunged so a bar of entry is needed to make it harder to abuse.

The point of contention here is mostly how high that bar should be. Low requirement makes it more accessible meaning we can spread the workload better but it'll be very vulnerable to abuse. High requirement would mean only somewhat experienced or rich people get access to it meaning much lower risk of abuse but very vulnerable to burnout. We'd have to find a sweet spot in the middle.

This post has been edited by Glovelove.: Apr 22 2022, 17:51

If it's a multiple of 1000, then any wrong ones are very unlikely to get reverted in a reasonable time frame, if ever.

Perhaps. Though considering said poster has basically been posting nothing but this for well over a year now I'm not sure if that's a great loss but I trust you guys can take that on balance and make a call on what makes sense.

Perhaps another approach you can take is not make it apply retro-actively in the first instance to see how people take it with new galleries for a while and then turn it on for older galleries once you're satisfied it doesn't cause problems?

(I apologise if I'm hogging the discussion on this, I still think this is a great, and probably well overdue change just sharing my 2 cents is all)

This post has been edited by Theonik: Apr 22 2022, 17:55

It is real pain after reading all the guidelines and knowing it is currently not okay to report people for scamming and adding ads onto works THAT THEY TOOK FROM EXISTING GALLERY.

If it's a dupe gallery + ads, it's handled through the duplicate expunge already, not this one.

If it isn't a dupe of a gallery, then having ads by itself won't qualify it for expunge, for example, ads at the end of the gallery are fine. But if a core image is modified to have an advert play as an animate gif, then that would come under defaced, or similar core content modifications for the sake of advertisements

And many translator tends to downsample the images not with malicious intention, but mainly because their available internet simply does not allow them to upload them in full size without constantly crashing.

Defaced (either option) doesn't cover such content, so no need to worry about that.

Was neutral on the mp requirement before, but starting to think a higher one would be better

I think the solution will also depend on what tenboro can provide as tools for moderation, which I can't really comment on because I have no clue how far those extend beyond the publicly available ones like tag history and expunge tracker.

If moderation could be provided with a way to check (recent) expunges and filter them by reason and uploader involved, and how many per uploader similar to how this allows us to see slaves per tag and click those to see the specific ones, moderators could spot an unexpected uploader name in there with a few uploads even if its 5 malicious expunges buried under 6000 cgc expunges and 500 more spread across a dozen of known abusers. On top of this it would aid in enforcing bans for spamming dupes/DNP, especially if there was an additional option to filter by recent only.

I know overviews like this should be possible from a programming standpoint, but the possibilities would still be limited by how many development time is available (especially on the short term) and whether Tenboro considers it worth that time, but the options to make this easier for moderators to keep up with are out there and ultimately those will decide how much freedom the community can have before the situation becomes too messy for mods to keep up with.

This post has been edited by Glovelove.: Apr 22 2022, 19:39

As far as any MP requirement, higher is better. If you are too low, then you can always post in the relevant thread if you spot an offending gallery. Same way you can for any tags that have a power threshold.

Better that way than a lower MP requirement more easily risking valid content getting expunged, even if it would be temporary, pending an appeal.

I'm glad for this.

Though the person who it'll primarily be targeting seems to have changed his tactics, probably as a result of this news...

Plastering sample over everything and uploading incomplete galleries with random pages.

So we're more or less back where we started.

*edit*

That said, instead of defaced (because that's slightly nebulous), how about Intentionally Vandalized?

This post has been edited by Jigsy: Apr 22 2022, 21:46

I was wondering why a Certain Gallery Creator had abruptly changed his modus operandi recently...