[X] My Assistant

We have been discussing adding a new expunge reason targeting works that have been intentionally defaced or degraded by the uploader or someone else in the chain. This was primarily triggered by recent galleries that had ads injected into the content, but it is also intended to address other undesirable behavior that results in content no one really wants.

Previously, expunging such galleries might have made them unlisted and undiscoverable, but after the last expunge and upload system revamps, expunged galleries can only be disowned, not deleted/unlisted, which makes this change feasible.

There are currently two competing definition for this expunge reason:

CODE

Content has been defaced by adding content-obstructing scanmarks, censorship or advertisements beyond what is present in the original artist release. (Scanmarks that do not obstruct content or promotional pages added after the content pages do not qualify.)"

-or-

CODE

"Content has been defaced by adding content-obstructing scanmarks, censorship or advertisements beyond what is present in the original artist release, or has been intentionally downsampled to the point where legibility is an issue. (Scanmarks that do not obstruct content or promotional pages added after the content pages do not qualify. Older low-resolution scans are exempt.)"

I'm leaning towards the former, simply because the threshold for "images for ants" will always be subjective and cannot be defined in hard numbers, but obviously, it will also make it not address some undesirable behavior.

I'd primarily want feedback on whether you think the "intentionally downsampled" clause will cause more problems than it solves, but feel free to discuss the addition in general as well.

What does this have to do with mtl/rough translation?

Generally the same people who'll take their raws from NH then get flamed for using downsampled resolution. If genuine incompetency is to be protected on that front it should get the same protection when it comes to resolution.

I think we all agree on that but I suspect people will get creative on where to draw the line of "legibility is an issue"

This post has been edited by Glovelove.: Apr 19 2022, 01:24

I can't imagine this effecting translations much at all. That site uses the same res as e-h's default, due to how they rip it. The only thing expunging those would do, is risk a ban.

I think we all agree on that but I suspect people will get creative on where to draw the line of "legibility is an issue"

I think we have enough active vig with eyeballs

Option 2 could be feasible because expunge tracker is a thing. Every actions and misbehaviors will be immediately reported to the community members, as opposed to something like tagging.

Option 2 could be feasible because expunge tracker is a thing. Every actions and misbehaviors will be immediately reported to the community members, as opposed to something like tagging.

Fair enough.

Alternative would be to set the bar low enough that 1280x can always get a "close enough" but this would probably invite more malicious compliance/trolling than it's worth

Alternative would be to set the bar low enough that 1280x can always get a "close enough" but this would probably invite more malicious compliance/trolling than it's worth

That's why we don't want to set a hard number

Just FWIW, it is of course possible to switch it later if the one we pick doesn't work out one way or the other.

Note that one of the reasons we don't want to set a target resolution is that there are plenty of ways to make a 1280x copy that looks like utter shit, say by downscaling it to a much lower resolution and then upscaling it again. As such, any sort of "you must be this wide to ride" kind of thing doesn't do much, it has to remain subjective to some extent.

Tbh, at this moment, it's kinda difficult for me to form an educated opinion as to whether option 1 or 2 will be better.

While "intentionally downsampled to the point where legibility is an issue" should be (or, I hope) a simple enough concept for folks with common sense to grasp, we also can't deny the possibility that the general community would be taking it too far. Again, with the expunge tracker, it is nothing that can't be countered. But obviously, we will need to revert that if the ruling proved extremely abusable. I would rather not camp the tracker all day just to fight the "rough resolution" mob 24/7.

I would say that it is perhaps worth giving option 2 a shot for now, if some community members call for it. But if the second one does end up causing a lot of issues, we can always revert the rulings back to option 1.

Yeah, subjectivity goes both ways, we'll just have to suffer from the options it leaves for abuse to reap the benefits it can provide to keep common sense around as a factor.

I think "intent" remains to be very troublesome concept for cases like this or the rewrite tag because it is nearly impossible to argue against if the uploader insists on having good intentions, but I don't think resolution fuzzy enough that it would get cryptozoologists excited is in anyone's interest. I'll support option 2 in that case and see where it goes, no objections to option 1 though, both cover the biggest problems perfectly fine.

This post has been edited by Glovelove.: Apr 19 2022, 03:15

Option 1 is a good addition. Option 2 is an extension of option 1.

Common sense is less common than we think, so I have no idea if the subjectivity that option 2 presents will be reliable without a more tangible threshold.

It's always possible to regress to option 1 if option 2 is proven abusive, so it's more of a question if it's worth trying option 2 or going straight to option 1.

My opinion is to try to make option 2 just a tad more palpable, but I really have no idea how (some suggestions were given before tho). If not possible, I prefer option 1, but wouldn't be opposed to option 2.

Option 2 suffers from the same problem of providing hard numbers by listing a single method of vandalism. If someone really wanted to, they could deepfry the images, increase contrast, or use a dozen other photoshop commands to deface a gallery beyond specifically downsampling it.
I would rather have the scope of vandalism methods to not be specified, as there are too many to list, than list new methods as trolls come up with new ways to shit up galleries.

As difficult as it may be to judge intent, a potential indicator is the uploader's history (cgc) or lack thereof (new spambots).

>Content has been defaced by adding content-obstructing scanmarks, censorship or advertisements beyond what is present in the original artist release, or has been intentionally edited (downsampled, blurred, pixelated, etc) to the point where legibility is an issue. (Scanmarks that do not obstruct content or promotional pages added after the content pages do not qualify. Older low-resolution scans are exempt.)

I would say that it is perhaps worth giving option 2 a shot for now, if some community members call for it. But if the second one does end up causing a lot of issues, we can always revert the rulings back to option 1.

I really do think this is the better way. Option 2 to me feels more correct and how this should be done if we're going to do it. And as pointed out, we can revert back to option 1 if it becomes an issue

If the subjectivity on 'intentionally down-sampled' is too much to leave to the masses, is there a way to limit it to more trusted users?
Tho I think for the most part this is fine anyway, it's rare for expunges to pass without the support of trusted users and their mod power.

This post has been edited by Cipher-kun: Apr 19 2022, 12:34

option 2 may put a bit more strain on the moderation team to check the validation of expunges? But on the other hand, everyone can start vote down any expunge petition. So if we have enough people on vigilante who can make a good enough estimate of what is still 'legible', option 2 could work.

I believe that most people act in good faith, and in that case option 2 would work fine. I'd say let's try option 2 first with fall-back to option 1 if needed.

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.

If the subjectivity on 'intentionally down-sampled' is too much to leave to the masses, is there a way to limit it to more trusted users?

^

Could consider adding a higher MP requirement to start or vote on a "defaced" petition. Unlike the other categories, it is a more subjective one after all.

Could consider adding a higher MP requirement to start or vote on a "defaced" petition. Unlike the other categories, it is a more subjective one after all.

Yeah, this would be trivial to do.

While it wouldn't prevent all attempted abuse of it, I'd imagine an MP requirement would significantly cut down on incorrect petitions being created. I'm in favor of implementing it.

I think providing hard numbers is a bad idea.

This seems like more of a discretionary expunge reason for annoying uploaders who sneak in ads, post "low quality" all over everything because the internet hurt them and people publishing tiny images.

Any hard number will just get abused.

With the new expunge type, do the existing rules apply about needing a higher quality copy to expunge? I've seen a couple uploaders deliberately posting the downsampled version of fanbox posts admitting so on their uploader comments. (example here) basically telling people to go to Fanbox for the full resolution images.

Would the suggestion be rather than replace and expunge these to simply expunge them and wait for a better version to arrive? The current rules broadly allow for the former through the replacement criteria. I'm assuming the intent here is to deter people from taking the piss deliberately challenging people to replace/expunge galleries and reaping the benefit while people do?

E: I'm broadly of the mind that the more broad the definition and flexible it is while remaining to the spirit of the rule the better so including the most broad set of common abuse (option two) makes the most sense in my mind, otherwise people will quickly abuse this, especially if hard numbers were added to the rule.

However as expunge reasons get more subjective I do think some more thought needs to be put into the mechanics of expunging and the incentives therein primarily three issues I think need to be addressed:
1) Remove the incentives to be the first to submit an expunge petition, and create infrastructure to allow more elaborate petitions to be created to explain the reasons for the expunge.
2) Create a way for makers of a petition to respond to downvotes on their petitions. If the argument is as I understand if whether it's better to have something or nothing, then people may downvote your petition, often times new information may surface that might require the amendment of the original petition but there is no way around this without cancelling and recreting the petition. Obviously editing a petition would be problematic, so letting people respond makes sense as a compromise as the strength of the evidence obviously needs to be proportionally higher in those cases.
3) Consider the punishment and creation mechanics for petitions on a gallery. Currently a failed petition carries many penalties. Those make sense where most petitions are fairly black and white and the main reason a petition might fail is a failure of reading the rules, these make less sense where petitions are a matter of opinion over relative value to the site which will likely be more controversial. (even with replacement petitions there can be controversy over what is better, which can disincentivise creating legitimate petitions to avoid risk on one's account)

N.B. I don't have exact answers for the above these are some crude suggestions and a matter of opinion. I'm sure a workable design could be worked out and these might prove uneccessary.

This post has been edited by Theonik: Apr 20 2022, 04:16

With the new expunge type, do the existing rules apply about needing a higher quality copy to expunge?

When this rule (either option) apply, it applies without requiring a replacement, though a replacement is of course beneficial to the community if one can be provided. Else, other expunge type rules apply. Any expunge that isn't defaced or forbidden, require a replacement or duplicate gallery linked to expunge.

I've seen a couple uploaders deliberately posting the downsampled version of fanbox posts admitting so on their uploader comments. (example here) basically telling people to go to Fanbox for the full resolution images.

Would the suggestion be rather than replace and expunge these to simply expunge them and wait for a better version to arrive?

No, it's not intended for cases like that gallery, they would be replaceable as normal.

When this rule (either option) apply, it applies without requiring a replacement, though a replacement is of course beneficial to the community if one can be provided. Else, other expunge type rules apply. Any expunge that isn't defaced or forbidden, require a replacement or duplicate gallery linked to expunge.

No, it's not intended for cases like that gallery, they would be replaceable as normal.

I guess the question kinda boils down to the grey area of what constitutes being defaced and what constitutes just being a poor job especially when it comes to things like poor quality scans/downsampled scans etc. This is somewhat subjective and as I understand it, the motivation is to deter people who deliberately deface works. (most of these could already be expunged but only via the replacement rule so there is a lot of overlap kind of inherently)

I guess the rule of thumb here is a combination of "is the poster taking the piss?" and "Is the community worse off having these than the alternative of having nothing?" That wasn't a perfect example I used back there, I guess my point was there is some nuance between the example I posted which is "defaced", but certainly not unreadable and many people would argue not expungible with the new rules, and what I suspect the motivation of the new rules is which is to fight the daily birthday spam of the "low quality" variety. (I.E. different people would place the line in a different place)