floral print tags?

BUR #22123 has been rejected.

mass update floral_print_pants -> floral_print print_pants -floral_print_pants
mass update floral_print_bow -> floral_print print_bow -floral_print_bow
mass update floral_print_shirt -> floral_print print_shirt -floral_print_shirt
mass update floral_print_apron -> floral_print print_apron -floral_print_apron
mass update floral_print_skirt -> floral_print print_skirt -floral_print_skirt
mass update floral_print_dress -> floral_print print_dress -floral_print_dress
mass update floral_print_bikini -> floral_print print_bikini -floral_print_bikini
mass update floral_print_kimono -> floral_print print_kimono -floral_print_kimono

I assume we don't want these.

If we really have to do it (which... please, let's not do it), don't for get to also add print_* tags, like print pants, to the BUR.

KSG said:

don't for get to also add print_* tags, like print pants, to the BUR.

Right, sorry. Updated it.

If we really have to do it (which... please, let's not do it),

Shall I assume you are volunteering to populate them then? Despite floral print being over 70k posts, these tags have a combined total of like 500, so even if we value them (which we tend to nuke tags like this), they don't come remotely close to doing a good enough job.

I've been populating some of those tags on the side on my own at my own pace, I have always maintained how the oldbooru custom of requiring 2-4 separate tags to tag a single is quite dumb, makes it impossible for blue users to find anything, completely breaks down if there is more of 1 character in the picture or if some tag padder starts tagging stuff in the background and leads to a whole lot of false positives (e.g. searthing for floral print + bikini gives you stuff like post# 2107478 just because the guy in the back is wearing a hawaiian shirt, or post #7042380 because she's wearing a floral print kimono in one of the views, or floral print + dress giving stuff like post #4256321 and post #3199428, and so on.

You can say that about pretty much any combination of tags on the site. Thus far, our policy on the matter has been to not want such tags. And if you're actually looking for, say, floral print bikinis, you search floral_print print_bikini, not floral_print + bikini, and that gives you substantially more results than floral print bikini does.

I'm not necessarily opposed to such tags but between how limited their utility is when they're this sparsely populated, the fact that it could be a never-ending game of "more granularity", and other topics like this one to get rid of frills (the umbrella tag) I am not so sure how it would go.

Talulah said:

You can say that about pretty much any combination of tags on the site. Thus far, our policy on the matter has been to not want such tags. And if you're actually looking for, say, floral print bikinis, you search floral_print print_bikini, not floral_print + bikini, and that gives you substantially more results than floral print bikini does.

That would require people to consistently tag both floral_print and print_[garment], just go check floral_print bikini -print_bikini and floral_print kimono -print_kimono

Mexiguy said:

That would require people to consistently tag both floral_print and print_[garment], just go check floral_print bikini -print_bikini and floral_print kimono -print_kimono

As opposed to floral_print_bikini/floral_print_kimono, which they are definitely tagging? Not tagging something is a problem regardless of whether you make more specific tags like this or not. All it changes is what tags you have to garden.

Here is some details on the frequency of some of these print objects and their depiction with one or more other print or patterned taggable articles of clothing.

Columns are:
(1) the tag name
(2) the percent or count of the tag being the only print_<object> present
(3) the percent or count of the posts depicting the print_<object> and other print/patterned objects sharing the same print
(4) the percent or count of the posts depicting the print_<object> and other print/patterned objects not sharing the same print
(5) Sample size of the random sample

Notes: When in doubt I defaulted to it counting toward single print object, such as not even seeing the tagged object, which happened a few times in print shirt, or there was another object with a potential pattern but one that likely wouldn't receive a tag.

Examples of the other objects that could be present in these images that might have a print beyond the one searched for were: handbags, underwear, ribbons, sashes, scarves, footwear, legwear, gloves, pants, and shorts.

Patterned objects were considered a different kind of print for the purposes of categorization. A checkered or plaid skirt for example was considered a different print from a character print shirt.

In cases of depictions of multiple characters, they were only counted as different prints if there was a different print/patterned object also present besides the searched for print_<object>. For example if there was an image of two characters in print kimono and both were different prints, if no other print objects were present then they would count toward the single print object category. If another print object was present, say a print handbag or print ribbon that only matched one of the character's print kimono, then it would be counted toward the Multiple Print Objects with a different Print category.

The Print Dress sample did not include Kisaki (blue archive) as she was making up a common repeat sample, though china_dress was still a major representation without her posts in the mix.

The Print Bikini sample did retain the cow print depictions, despite their very heavy representation in the sample. The largest amount of depictions with multiple articles of clothing sharing the same print were due to cow print images featuring a cow print bikini with matching cow print thighhighs and gloves. tiger print had a lesser but similar impact as cow print to the results.

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

This bulk update request has been rejected because it was not approved within 45 days.

The bulk update request #22123 (forum #263748) has been rejected by @DanbooruBot.