Danbooru

Pixiv 1000 Question

Posted under General

Magus said:

I see no reason to get rid of it.

Reasons have already been explained in this thread. If you don't see them then you haven't read the thread.

If you're going to argue for the pool to stay then explain why.

Toks said:

Reasons have already been explained in this thread. If you don't see them then you haven't read the thread.

If you're going to argue for the pool to stay then explain why.

Or perhaps I just don't see your reasons as compelling.

Actually, it would be interesting to see the pictures which piqued pixiv dwellers' interest pooled together, without going out of my way to actually search the most bookmarked images on pixiv and then to try to find those on danbooru. But 1000 just catches a lot of junk now - that's why I proposed to raise the plank to 10000 (or higher?) and re-populate the pool.

The pool is too meta, without any common thread in the contents themselves. Moving the goalposts isn't going to change that. Tag Pool what you see. How many favorites a given post supposedly has on a different site--even pixiv--is of no consequence to ourselves.

And anyway, who watches the watchmen? I count 72 posts in that pool with less than 1000 pixiv favorites, and at least 1885 posts that can't be verified.

Below 1000 favs

post #1473405: 825 favs
post #1456069: 29 favs
post #1450921: 834 favs
post #1442528: 647 favs
post #1419768: 173 favs
post #1395042: 786 favs
post #1357331: 583 favs
post #1342047: 580 favs
post #1329525: 332 favs
post #1320973: 360 favs
post #1315758: 849 favs
post #1312053: 585 favs
post #1280140: 925 favs
post #1278917: 809 favs
post #1268054: 874 favs
post #1268047: 436 favs
post #1210391: 391 favs
post #1178761: 733 favs
post #1161557: 980 favs
post #1159390: 573 favs
post #1154287: 878 favs
post #1153166: 370 favs
post #1151633: 68 favs
post #1112152: 930 favs
post #1091804: 194 favs
post #1089858: 388 favs
post #1053978: 660 favs
post #1053444: 922 favs
post #1041498: 726 favs
post #943395: 634 favs
post #937447: 543 favs
post #925081: 154 favs
post #924617: 338 favs
post #923271: 347 favs
post #922095: 633 favs
post #921122: 560 favs
post #916488: 473 favs
post #916319: 679 favs
post #908763: 5 favs
post #901224: 197 favs
post #896656: 246 favs
post #896606: 513 favs
post #894605: 306 favs
post #892826: 580 favs
post #882688: 980 favs
post #857439: 794 favs
post #841811: 624 favs
post #838700: 630 favs
post #834406: 497 favs
post #830695: 850 favs
post #826949: 131 favs
post #814664: 392 favs
post #774002: 844 favs
post #762889: 808 favs
post #762888: 808 favs
post #762884: 808 favs
post #762081: 567 favs
post #750887: 475 favs
post #728856: 919 favs
post #715422: 174 favs
post #709235: 960 favs
post #674746: 455 favs
post #673786: 446 favs
post #635333: 315 favs
post #633192: 144 favs
post #613652: 309 favs
post #597434: 970 favs
post #472847: 766 favs
post #443010: 560 favs
post #400322: 985 favs
post #341347: 463 favs
post #251196: 903 favs

Final distribution:

  • [0,1k) = 72
  • [1k,5k) = 11395
  • [5k,10k) = 1899
  • [10k,+) = 828

Delete.

It's still unclear to me what there is to gain by deleting the pool.
If someone doesn't find value in the pool, I'm not sure how its presence is causing them a problem.

who watches the watchmen? I count 72 posts in that pool with less than 1000 pixiv favorites, and at least 1885 posts that can't be verified.

Well, you've demonstrated RaisingK can watch the watchmen. In fact, what you've done isn't even possible with most pools--you can run an algorithm and determine which posts shouldn't be in the pool.
The error rate of verifiable posts is 72/(72+11395+1899+828-1885)= 0.00584. That seems like a low enough margin that it's not really an issue. I imagine the error rate for the unverifiable posts is similar.

It's still unclear to me what there is to gain by deleting the pool.
If someone doesn't find value in the pool, I'm not sure how its presence is causing them a problem.

Same could be said of all the other pools we delete? We do have standards here.

Well, you've demonstrated RaisingK can watch the watchmen. In fact, what you've done isn't even possible with most pools--you can run an algorithm and determine which posts shouldn't be in the pool.

Indeed, but I'm not exactly interested in running anything like this pool.

RaisingK said:
Sets a bad example. Not as bad as the Perfect pools, but still.
Indeed, but I'm not exactly interested in running anything like this pool.

Just being a "bad example" isn't that compelling a reason to me. Has it actually inspired any significant bad behavior/results? How has the presence of the pool caused someone significant problems?
I didn't intend to volunteer you for watching, but just that it is possible. Even without watching it has a very low error rate.

Some people have proposed converting the pool to a tag. I'd be fine with either. If we were to convert the pool to a tag, someone would just have to make sure to move all the pool's current info into the tag's wiki article.

Some people have proposed raising the bar from 1000 to 5000 or 10000. I'd prefer to keep the 1000 and simply add an additional bar. Imagining a scenario where we had two bars like 1000 and 5000, posts put in the 5000 bar we'd want to automatically also place in the 1000 bar. I don't know if that's possible with pools? I know it would be with tags and tag implications. (Is it possible for pool:X to imply pool:Y?)

As it stands, 1000 isn't really the milestone it used to be. A bigger problem is what this information serves. Its a nice way to bloat pool numbers, and given that we nuked the "perfect" pools, there is plenty of precedent to nuke it for being a "bad example."

The issue with pixiv 1000 is that we don't have equivalent tags for any other source for our images (granted pixiv is the primary source, and most others don't have a way to track information).

Changing the goal post serves nothing since its a lot of work to go back and update the information, and it doesn't make it any more useful.

On the flip side, Its a nice point of trivia, and it can be used for finding good quality comics/art that don't have a high score/favcount.

Saladofstones said:
On the flip side, Its a nice point of trivia, and it can be used for finding good quality comics/art that don't have a high score/favcount.

If we were really interested in what other sites' communities think about images, we'd copy and add scores from all other boorus to our posts.
But we aren't. Even our normal members cannot even vote.

Saladofstones said:
On the flip side, Its a nice point of trivia, and it can be used for finding good quality comics/art that don't have a high score/favcount.

True.

S1eth said:
If we were really interested in what other sites' communities think about images, we'd copy and add scores from all other boorus to our posts.
But we aren't. Even our normal members cannot even vote.

Also true.

It's nice to have but I guess it can go. If you look for good pictures you can sort by score or favs. If you're interested in some other site's ratings you can look at that site.

1 2 3 4 5 6