Deprecating frills

BUR #22083 has been rejected.

deprecate frills
mass update frills ~frilled_* ~center_frills ~gathers ~too_many_frills -> -frills

While it might seem like a good idea to have these kinds of parent tags, frills' existence has only enabled lazy uploaders and caused severe undertagging - we currently have over 80,000 posts tagged with just frills. Almost every single post in the first hundred results for frills -frilled_* -center_frills -gathers -too_many_frills has a clear and obvious frilled_x tag that should have been used in its stead, but (I assume) because frills is so easy to type, the uploader chose the path of least resistance and didn't bother.

This BUR aims to combat such laziness by removing that path of least resistance. If you feel like downvoting it because some kinds of frilled clothing don't have a frilled_x tag, then make them. We already have almost forty frilled_x tags, and frills alone does not actually help at all - if, say, frilled_footwear didn't exist, then I'm sure you would have a grand old time using the frills shoes search and sorting through almost two thousand pages of false positives.

"But Zapdos!" I hear you cry. "However will I search for frills in general if not for frills?" Well, *frill* is a 99.999% identical search that actually performs even better than frills, because it catches all posts with frill tags regardless of whether they imply frills or not. Yes, such a search will catch some posts that have nothing to do with frills. However, given that there are only 245 of those, the almost 400,000 genuine *frill* posts make it so you'll only see one false positive every 83 pages. I think even our most anal of users can handle that.

Is there any reason this applies to frills only and not the plethora of other similar umbrella tags? Things like strapless and fur_trim come to mind.

Talulah said:

Is there any reason this applies to frills only and not the plethora of other similar umbrella tags? Things like strapless and fur_trim come to mind.

One at a time. A BUR trying to deprecate all of these at once would be doomed from the start.

This bulk update request is pending automatic rejection in 5 days.

What comes to mind to solve the issue of users not tagging specific frilled_*, while still allowing blanket frills searches, is the idea of "virtual" tags.
Like deprecated tags, virtual tags can't be added by users manually, but, unlike deprecations, may still be added through implications, and are automatically deleted if no antecedent tag is present.

hdk5 said:

What comes to mind to solve the issue of users not tagging specific frilled_*, while still allowing blanket frills searches, is the idea of "virtual" tags.
Like deprecated tags, virtual tags can't be added by users manually, but, unlike deprecations, may still be added through implications, and are automatically deleted if no antecedent tag is present.

That sounds like a good idea, and much better than just deprecating stuff like this.

As I mentioned in my initial post, *frill* is already and will still be a catch-all search for anything frills if this BUR is approved. The desire for it to be a tag rather than a wildcard search is not nearly enough of a reason to keep frills around when it's enabling so much lazy tagging.

frills is a perfectly good tag that is far more intuitive than throwing wildcards around.

Throwing away a perfectly good tag just because it might be causing underuse of a bunch of more obscure and specific tags seems foolish.

skylightcrystal said:

frills is a perfectly good tag that is far more intuitive than throwing wildcards around.

Throwing away a perfectly good tag just because it might be causing underuse of a bunch of more obscure and specific tags seems foolish.

What would frills catch that frilled_* wouldn't? Seeing as other broad tags like striped got depreciated in favor of striped_* this BUR makes sense.

zetsubousensei said:

What would frills catch that frilled_* wouldn't?

Things like post #6234977 that have frilly designs, but no applicable "frilled x" tags, even if you tell someone to "just make it". How exactly would you tag the frilly parts on this one beyond just frills?

Maiden_in_Orange said:

Things like post #6234977 that have frilly designs, but no applicable "frilled x" tags, even if you tell someone to "just make it". How exactly would you tag the frilly parts on this one beyond just frills?

Ignoring the madness that is tagging animals' fins with frills, you seem to be missing the crux of the deprecation. The question is not "how will I tag this if frills is deprecated?", the question is "how would I use frills to search for this in the first place?". There's only a little over 100 posts under animal_focus frills, but even with such a small sample size you can still see the problem - finding frilly fins without a tag specifically for them requires you to sift through over a hundred posts of animals in frilly clothes. On its own, frills is completely useless because of the insanely high rate of false positives when trying to find literally anything without its own frilled_x tag, and that's the other main reason I'm trying to deprecate it.

All that being said, since frills is for the man-made clothing adornments, I've created the animal_frills tag. As I've hopefully made very clear, it's a huge chore to search for them without such a tag (which unfortunately also means I've only been able to find three posts to add it to).

Frills as a tag was quite useful for me when populating jersey maid. A huge OR search for every frill-related tag would have been annoying to conclude.

zetsubousensei said:

What would frills catch that frilled_* wouldn't? Seeing as other broad tags like striped got depreciated in favor of striped_* this BUR makes sense.

frills are embellishments. stripes are patterns. false analogy.

While yes, I agree that frills can absolutely be a complete dumping ground for lazy taggers unwilling to tag a frilled skirt properly (which can be incredibly frustrating), I disagree that outright deprecating the tag entirely is a great idea. And not just from a tagging standpoint.

People trying to find literally anything with frills are not going to intuitively search frill* or frilled_*. They're going to search the tag frills, which automatically shows up in the search bar, and gives them frills for days. Do you really expect people to immediately resort to wildcards the moment that tag becomes unavaliable? A search that is far less intuitive and does not automatically show up in the search bar, meaning the average casual member level user with no search knowledge outside basic tags is kinda out of luck. And we can only redirect people to the Cheat Sheet or something when they come wanting to undeprecate this tag due to not being able to search that so many times before it starts getting annoying. Expecting them to without being informed Danbooru supports wildcards sounds completely asinine. And we all know people have a bad tendency to not read the damn wikis. The fact that there are completely avoidable false positives too in a frill* search compared to simply using frills just raises more alarm bells, no matter how few they are all around.

Now admittedly, I don't have any better solution for the mintagging conundrum besides "just garden it", which literally no one wants to hear (I certainly wouldn't). So I'm not going to seriously suggest that. Nor do I think there are too many people that are seriously going "I want to see literally anything with frills on it!" on a regular basis. But what if I wanted to see, say, Hakurei Reimu without all the frills? Or what if I wanted to see any given character normally lacking in frills (let's go with Luigi here because why not) in a frilly outfit? Or KSG's example in populating the jersey maid tag? Frills already covers my basis compared to wildcard searches. And shouldn't that be taken into consideration at least? Going through with an extreme motion like full-on deprecating the tag would sacrifice searchability for the people with searches like these, especially when the proposed new method of searching is far less intuitive than the tag on the chopping block.

Maiden_in_Orange said:

People trying to find literally anything with frills are not going to intuitively search frill* or frilled_*. They're going to search the tag frills, which automatically shows up in the search bar, and gives them frills for days. Do you really expect people to immediately resort to wildcards the moment that tag becomes unavaliable? A search that is far less intuitive and does not automatically show up in the search bar, meaning the average casual member level user with no search knowledge outside basic tags is kinda out of luck. And we can only redirect people to the Cheat Sheet or something when they come wanting to undeprecate this tag due to not being able to search that so many times before it starts getting annoying. Expecting them to without being informed Danbooru supports wildcards sounds completely asinine. And we all know people have a bad tendency to not read the damn wikis.

Searching for an empty tag brings up its wiki, which can very easily be rewritten to include something along the lines of "To search for all kinds of frills, use *frill*.", which is simple enough that anyone who would actually be interested in searching for it is likely going to understand what to do.

Maiden_in_Orange said:

The fact that there are completely avoidable false positives too in a frill* search compared to simply using frills just raises more alarm bells, no matter how few they are all around.

0.06% of the results are false positives. As mentioned before, that's a minuscule 245 out of about 400,000 posts, which is around one false positive every 83 pages of results. You'd have to be scrolling through more than one page a minute just to find one false positive every hour; nobody is going to complain about this.

Maiden_in_Orange said:

Frills already covers my basis compared to wildcard searches. And shouldn't that be taken into consideration at least? Going through with an extreme motion like full-on deprecating the tag would sacrifice searchability for the people with searches like these, especially when the proposed new method of searching is far less intuitive than the tag on the chopping block.

Keeping the tag around impacts searchability in a different but more impactful way. Compared to the users that even you admit are relatively few and far between who are just searching frills, there are many more searching for the individual subtags like frilled_skirt. These users are having their searches negatively impacted by frills' mere existence in it enabling lazy tagging.

I don't think "enabling" is the right word. If someone doesn't tag frilled skirt, only frills, I don't think they'll magically start tagging frilled skirt just because their go-to tag stopped existing, they just wouldn't tag the concept at all.

KagayakuShiningGate said:

I don't think "enabling" is the right word. If someone doesn't tag frilled skirt, only frills, I don't think they'll magically start tagging frilled skirt just because their go-to tag stopped existing, they just wouldn't tag the concept at all.

Isn't that what happened when tags like the colored *_legwear tags got removed? People started using more specific tags like socks or thighhighs. Maybe it won't magically change things overnight but it's a good nudge. I like the suggestion Zapdos said to update the wiki to explain how to use *frill*

AngryZapdos said:

BUR #22083 has been rejected.

deprecate frills
mass update frills ~frilled_* ~center_frills ~gathers ~too_many_frills -> -frills

While it might seem like a good idea to have these kinds of parent tags, frills' existence has only enabled lazy uploaders and caused severe undertagging - we currently have over 80,000 posts tagged with just frills. Almost every single post in the first hundred results for frills -frilled_* -center_frills -gathers -too_many_frills has a clear and obvious frilled_x tag that should have been used in its stead, but (I assume) because frills is so easy to type, the uploader chose the path of least resistance and didn't bother.

This BUR aims to combat such laziness by removing that path of least resistance. If you feel like downvoting it because some kinds of frilled clothing don't have a frilled_x tag, then make them. We already have almost forty frilled_x tags, and frills alone does not actually help at all - if, say, frilled_footwear didn't exist, then I'm sure you would have a grand old time using the frills shoes search and sorting through almost two thousand pages of false positives.

"But Zapdos!" I hear you cry. "However will I search for frills in general if not for frills?" Well, *frill* is a 99.999% identical search that actually performs even better than frills, because it catches all posts with frill tags regardless of whether they imply frills or not. Yes, such a search will catch some posts that have nothing to do with frills. However, given that there are only 245 of those, the almost 400,000 genuine *frill* posts make it so you'll only see one false positive every 83 pages. I think even our most anal of users can handle that.

Then the correct way to handle this is tag things up, and not remove the tag completely.
We'd be making tagging less precise if this were to go through.

zetsubousensei said:

Isn't that what happened when tags like the colored *_legwear tags got removed? People started using more specific tags like socks or thighhighs. Maybe it won't magically change things overnight but it's a good nudge. I like the suggestion Zapdos said to update the wiki to explain how to use *frill*

Black_thighhighs is ultimately less typing than black_legwear thighhighs. Additionally, people made that jump more easily as there were no black_thighhighs/pantyhose/socks tags before to imply black_legwear; the change was like, 1:1 in the amount of effort a lazy tagger needed to put in. Frills as a tag is an umbrella with subtype tags.

KagayakuShiningGate said:
I don't think "enabling" is the right word. If someone doesn't tag frilled skirt, only frills, I don't think they'll magically start tagging frilled skirt just because their go-to tag stopped existing, they just wouldn't tag the concept at all.

...

Black_thighhighs is ultimately less typing than black_legwear thighhighs. Additionally, people made that jump more easily as there were no black_thighhighs/pantyhose/socks tags before to imply black_legwear; the change was like, 1:1 in the amount of effort a lazy tagger needed to put in. Frills as a tag is an umbrella with subtype tags.

The legwear tags were brought up because they are proof that, rather than abandoning all efforts to tag them, users will switch to better tags when the inferior one(s) they've been using are deprecated or otherwise made obsolete. To assume the opposite will hold true this time is to completely ignore this very large case study - should we just let this issue fester on the off chance that (hypothetically) this goes through and the users who only tag frills all just decide in unison to not start using the subtags instead? All for the sake of a tag whose function is replicable with the simplest of wildcard searches? This BUR hasn't even been approved, yet you're speaking as if this defeatist scenario (which flies in the face of all available evidence) is a foretold conclusion.