Replacing posts with higher res versions

The problem I’m about to describe may seem trivial to some, but it really annoys me deeply. I’d like to propose a fix for it.

One thing that happens all too often here is that someone will upload a really great picture, and then a few months later someone will discover the same picture but in higher resolution. Sure we have the parenting system to link the old to the new, but that doesn’t carry over tags, notes, score, favorites, comments, etc. Anyone who has the original favorited has to manually remove that one and add the new one if they want to stay current. If the picture has notes, they must all be manually recreated.

Another problem is that for especially popular pictures, it seems to be the case that the new image will rarely surpass the original in score, simply because the people who would have up-voted it feel they’ve already done so. Thus if someone searches order:score, the original usually show up first, leading to more people faving it/up-voting it rather than the better version, exacerbating the problem. (For example, see post #530590 vs post #676605). The end result is that you get scenarios like “this version has higher res, but this version is higher ranked/translated/better tagged/has funnier comments/whatever” which should not arise.

I think the ideal way to fix this is to create a way to actually replace posts with higher-resolution versions, keeping all the post’s attributes intact. The ability to merge two existing posts is also a solution, although then you have to worry about issues like how to merge their score, if the same person voted up both. But what I’m suggesting is the ability to an image as a replacement for another, right off the bat. Obviously, this would be something that would have to go through moderation somehow, because it’s rife with opportunities for abuse. So, yes, more work for the janitors. And yes, it’s a major change. But I think it would really help make Danbooru a little more organized. Thoughts?

No.

I honestly don't see a good reason for this idea. Just because an image is higher-res does not mean that it's superior. It happens a lot where a high-res scan is uploaded, but the smaller image uploaded earlier is much clearer.

If the new image is parented to the smaller image then there is no problems with the fact that more people might have favorited it over the newer image. The newer image is easily accessible to those who have favorited the parent by looking at the child posts, so if there are also other variants out there they'll have access to those as well. I don't see it as some burden that you have to manually remove one from your favorites and add the other. If someone doesn't bother looking at parent/child posts and favorite a "inferior" image, then obviously they don't care, as that image was good enough for them to favorite it to begin with.

The fact that one his a higher score than the other is irrelevant, scores should hold very little weight on judging an image. While large scores might mean that an image is good, it doesn't really mean much. All it indicates is a slightly greater probability that it might be good, but even then I honestly believe that it's only a small percent greater in probability.

I don't see a great reason comments should be carried over from one to the other. Notes though would be nice, but at the same time it's not hard for someone to cut and paste notes from one to the other. Additionally you get times where people will do a new translation with the newer one, which might seem like wasted time, but I think it's more of a positive thing to see how others would have translated the same text.

It might be an interesting idea though to consider for identical images that one could make it so that they share the same comments thread for all of them, but overall it's need is minor, as it'll be infrequently used and more often used poorly. Most long comment threads aren't all that positive in how they're utilized.

Agreed, no.

Higher resolution is not intrinsically "better". It may be for some, but others might prefer the moderate resolution. Don't make that choice for them. Implementing something like this will lead to many arguments about which version is superior. Just like when you have a decensored pic (maybe at lower res), and the original pic. Which is better? Higher res comes with longer loading times, which is a disadvantage even if the resolution is the only thing that differs. And when a pic has replaced an older pic, how do you prevent someone from uploading the older pic again and replacing this one? It will also open a lot of new doors for vandalism.

Score doesn't mean anything. Tags can be copied trivially anyway.

Favorites? Maybe he intentionally wants to favorite the lower resolution anyway. Otherwise, when he looks at the pic again, he'll note that it has a parent and then make a decision to parent that or not. And if he doesn't look at the pic again, what difference does it make?

Linking comments somehow, it might be nice, but almost certainly isn't worth the trouble to implement it.

Copying notes, now that's a feature worth talking about perhaps.

I'm not arguing that all images with higher res available should be replaced. Nor am I arguing that parenting should be abolished in all cases. However, there are many images of this nature, like the pair I pointed out, where one is just unquestionably better than the other, and the lower res version exists just to suck up votes/favs/bumps that should go to the parent. Those are the images I'd like to see addressed.

I guess the favoriting issue only applies to weirdos like me who are aghast at the idea that they may have favorited an inferior version of an image, but I disagree that the score issue is irrelevant. As you say, score is only a probabilistic measure of image quality, but the fact is that it's one of the only ways we have to find good images in this sea of artwork. That makes it important.

And as for translations, it's usually far more work than just copy/pasting. Especially for images with many speech bubbles, positioning the notes is what takes up time.

I'll agree that it's certainly not our most pressing issue, but I think it could be a useful feature, if done correctly.

I understand what you're saying. I think the main obstacle here is that people won't see it as enough of a problem to be worth fixing.

If it wouldn't be too much trouble then I'd also like to see this get fixed

you should probably edit the title to "better quality" instead of "higher res"

"higher ranked/translated/better tagged/has funnier comments/whatever"
The above quote doesn't matter aside from translation, and doesn't have any impact on our handling of variant version of a post.

There are probably interesting things we could do with automating comment and tag and note carryover (when they're relevant to the new image, they wont always all be), but it would be pretty far down on my own priority list, to be honest.

Parenting seems like a perfectly good working alternative for the moment.