Topic: Regarding "cub" tag removal and why this is the worst decision in this site's history.

zombieastronaut said:
I highly doubt a user that search cubs there wants to see human lolis/shotas or teens but from now on they have to.

Just blacklist the tags you do not want to see.

zombieastronaut said:
Since there is no way to search for cubs and cubs only safe for blacklisting/excluding all tags related to humans and teens.

young -human

aaronfranke said:
young -human

young_human exists for exactly this purpose, same with young_humanoid. In my blacklist, I use this in order to get a similar effect to the old cub tag:

young_human -rating:safe
young_humanoid -rating:safe

kyureki said:
young_human exists for exactly this purpose, same with young_humanoid. In my blacklist, I use this in order to get a similar effect to the old cub tag:

young_human -rating:safe
young_humanoid -rating:safe

cub wasn't only for -rating:s posts.

I'm going to be honest, cub wasn't ever used properly, it really always was just the young tag but less tagged. I went through and manually retagged 8000 posts from set:cub_apocalypse, there was absolutely no consistency for what it got used on; humans and humanoids were tagged with it pretty much just as often as anthros and ferals.

dba_afish said:
cub wasn't only for -rating:s posts.

I'm going to be honest, cub wasn't ever used properly, it really always was just the young tag but less tagged. I went through and manually retagged 8000 posts from set:cub_apocalypse, there was absolutely no consistency for what it got used on; humans and humanoids were tagged with it pretty much just as often as anthros and ferals.

Agreed that there was never a solid definition, hence why I am glad it's gone. I am much happier with the young_* tags, as it allows me to better tune my blacklist (as described above) how I'd like.

dba_afish said:
I went through and manually retagged 8000 posts from set:cub_apocalypse, there was absolutely no consistency for what it got used on; humans and humanoids were tagged with it pretty much just as often as anthros and ferals.

This isn't true, to the point of taking up a really bad faith position. The last batch of the set that you dealt with was the worst of it, which can give a false impression.

set:cub_apocalypse limit:250 young_anthro - almost 123K
set:cub_apocalypse young_feral - almost 20K
set:cub_apocalypse young_human -young_anthro -young_feral - ~350
set:cub_apocalypse young_humanoid -young_anthro -young_feral - ~1700

Like I said when the changeover was proposed, before anything was actually done, cub was actually tagged way more consistently than most of our other huge tags. And now the numbers back that up conclusively. The mistag rate was less than 2%, which is really good for such a supposedly awful tag. Big "told you so" moment for me.

Once again, the main problem with the cub tag was a loss of critical thinking when people saw it. Huge negative connotation and inability to reconcile that 3-4' characters look like children. I'm not sure if that's lessened with the changeover or if it's same shit, different pile. And, as it's turned out, many people tagging and searching cub only wanted it on "definitely children," maybe subconsciously capping the age limit to 15 years instead of all non-adult furry characters, which does actually reinforce consistent tagging, just not to the standard that the website requires. Furthermore, by now I've seen many, many people admit that they have no idea how to tag age ranges, so it comes as no surprise that adolescent/teenager is a hot mess of lore tagging and that cub taggers avoided that full age group.

People are still tagging and searching cub and whining in the comments when they get young humans and humanoids instead, specifically saying they searched "cub" or that the post "isn't cub," even though the tag has been dead for months.

abadbird said:
This isn't true, to the point of taking up a really bad faith position. The last batch of the set that you dealt with was the worst of it, which can give a false impression.

set:cub_apocalypse limit:250 young_anthro - almost 123K
set:cub_apocalypse young_feral - almost 20K
set:cub_apocalypse young_human -young_anthro -young_feral - ~350
set:cub_apocalypse young_humanoid -young_anthro -young_feral - ~1700

sorry, what I meant to say was that the ratio of each young form within cub was roughly similar to the amount in young itself. the overall ratio of young anthro, feral, humanoid, human is roughly 48:8:4:3. while the ratio of the same within cub was 48:8:1:1.

so, I guess I am still pretty wrong here, humans and humanoids made up about 3.5% of cub vs 9% of young. I'm not sure if this was just my perception or maybe there was a higher than average amount of human and humanoid posts in the 8000 that needed to be manually handled.

dba_afish said:
I'm not sure if this was just my perception or maybe there was a higher than average amount of human and humanoid posts in the 8000 that needed to be manually handled.

The amount of human(oids) in what was leftover would have been a large concentration considering I did the bulk of the scripting and I only did anthro & feral since that was the safest to automate
Whatever was leftover would have either been ambiguously tagged, or human(oid)

I don't know how to search for what I want anymore. Young is such a broad ranging term, it makes finding what I want practically impossible. Not even specifically talking about cub, but all the different terms being consolidated into young makes it impossible to differentiate characters that are smaller and younger than others. If you want to look for art of characters in their teens specifically that's impossible now. If you want to find characters that appear younger than that specifically that's impossible now. I understand there's a level of subjectivity but that should come to defining the terms for the site, not consolidating a bunch of different terms together. Just earlier I searched "cub" and got a picture of a father/son pairing where the son was tall and had muscle tone, but because he's not an adult he showed up in that tag when that's not what I was looking for. And I can't even easily separate it out! Its trading subjective vagueness for overly broadness.

At the very least tell me how I can find characters that look more like something Dagasi would make rather than something...older looking?

meowpurryiff said:
I don't know how to search for what I want anymore. Young is such a broad ranging term, it makes finding what I want practically impossible. Not even specifically talking about cub, but all the different terms being consolidated into young makes it impossible to differentiate characters that are smaller and younger than others. If you want to look for art of characters in their teens specifically that's impossible now. If you want to find characters that appear younger than that specifically that's impossible now. I understand there's a level of subjectivity but that should come to defining the terms for the site, not consolidating a bunch of different terms together. Just earlier I searched "cub" and got a picture of a father/son pairing where the son was tall and had muscle tone, but because he's not an adult he showed up in that tag when that's not what I was looking for. And I can't even easily separate it out! Its trading subjective vagueness for overly broadness.

At the very least tell me how I can find characters that look more like something Dagasi would make rather than something...older looking?

child

meowpurryiff said:
(...)

We only merged cub into young. We didn't touch anything else. Cub had no age range, it only ever appeared as such because it was widely mistaken. A search for ~young_anthro ~young_feral should approximate how the cub tag should have been.

dba_afish said:
child

wow okay that's exactly what I was looking for

donovan_dmc said:
We only merged cub into young. We didn't touch anything else. Cub had no age range, it only ever appeared as such because it was widely mistaken. A search for ~young_anthro ~young_feral should approximate how the cub tag should have been.

Thank you for correcting me! I thought several things were merged into young (loli, shota, cub, kemoshota, and teen). I can see that I was mistaken, and I also see that there are tags to help get what I'm looking for. I apologize. Thank you!

meowpurryiff said:
I don't know how to search for what I want anymore. Young is such a broad ranging term, it makes finding what I want practically impossible. Not even specifically talking about cub, but all the different terms being consolidated into young makes it impossible to differentiate characters that are smaller and younger than others. If you want to look for art of characters in their teens specifically that's impossible now. If you want to find characters that appear younger than that specifically that's impossible now. I understand there's a level of subjectivity but that should come to defining the terms for the site, not consolidating a bunch of different terms together. Just earlier I searched "cub" and got a picture of a father/son pairing where the son was tall and had muscle tone, but because he's not an adult he showed up in that tag when that's not what I was looking for. And I can't even easily separate it out! Its trading subjective vagueness for overly broadness.

At the very least tell me how I can find characters that look more like something Dagasi would make rather than something...older looking?

In addition to what the others said, you might want to take a look at the young wiki page. It contains helpful sections listing out both the different age-based tags and the different form-based tags. You can usually find the wiki for a tag by clicking the "?" next to a tag, and quite often they contain useful info about sub-tags.

dimoretpinel said:
I already explained to you that you can have the same results searching only young if you use your blacklist.

You need to add to your blacklist the tags young_human, young_humanoid and adolescent.

The blacklist will weed out the undesired posts from your usual young search, effectively turning the young tag into your definition of "cub".

https://e621.net/help/blacklist

I think this might be necro but you cant really be suggesting its easier to make a blacklist and constantly click on and off its contents along with typing a longer word is any easier than typing in cub. I feel like the site is losing its identity as it changes the tag systems into more and more generic lists that just make things harder to find like how i cant find dragonesses anymore due to the tag aliasing into something thats not a dragoness.

tester29 said:
I think this might be necro but you cant really be suggesting its easier to make a blacklist and constantly click on and off its contents along with typing a longer word is any easier than typing in cub. I feel like the site is losing its identity as it changes the tag systems into more and more generic lists that just make things harder to find like how i cant find dragonesses anymore due to the tag aliasing into something thats not a dragoness.

My guy, this change is a done deal
It has been some 18 months since it happened

If taking poorly defined combination terms and splitting them into a few more well defined terms is making the site lose its identity, I'm not really sure it needs that identity

donovan_dmc said:
I will reiterate again that NOTHING would be approved if it needed to involve any percentage of site users. Most users simply do not care. We'd never get anything done if we had some arbitrary vote threshold to get something accepted.

But then why did the other got an announcment and furthermore we got that april fools prank where everyone had to draw their fursonas as an ID, if that is possible dont tell us it would been hard to put a screen up for every user logged in that says "We are aliasing cub with young. Do you agree yes/no?"

donovan_dmc said:
My guy, this change is a done deal
It has been some 18 months since it happened

If taking poorly defined combination terms and splitting them into a few more well defined terms is making the site lose its identity, I'm not really sure it needs that identity

I know its done now, this wont change anything. I just noticed it today that the alias changed into young and it annoys me that something that part of the fandom was changed to generic tagging thats just as vague on what it means but now its also soulless.

tester29 said:
I know its done now, this wont change anything. I just noticed it today that the alias changed into young and it annoys me that something that part of the fandom was changed to generic tagging thats just as vague on what it means but now its also soulless.

what kind of soul is in "cub"

tester29 said:
But then why did the other got an announcment and furthermore we got that april fools prank where everyone had to draw their fursonas as an ID, if that is possible dont tell us it would been hard to put a screen up for every user logged in that says "We are aliasing cub with young. Do you agree yes/no?"

Aliasing tags is hardly ever worth a full site announcement

Not to mention we've never seen a full scale test of voting, from what I've seen it wouldn't go well and would likely crash and burn due to numerous performance issues

Also, if you read the multitudes of topics surrounding the change you'd realize there was some kind of announcement planned which subsequently got canceled when it was overshadowed by the site getting banned in North Carolina

tester29 said:
[...]more generic lists that just make things harder to find like how i cant find dragonesses anymore due to the tag aliasing into something thats not a dragoness.

my dude, dragoness has been aliased away for over a decade, and, when it was active, it had only been added to a total of 100 posts. you don't know what you're talking about.

tester29 said:
I think this might be necro but you cant really be suggesting its easier to make a blacklist and constantly click on and off its contents along with typing a longer word is any easier than typing in cub. I feel like the site is losing its identity as it changes the tag systems into more and more generic lists that just make things harder to find like how i cant find dragonesses anymore due to the tag aliasing into something thats not a dragoness.

Reply to this post with the exact search query you want to use, because you're almost certainly doing something wrong.

tester29 said:
But then why did the other got an announcment and furthermore we got that april fools prank where everyone had to draw their fursonas as an ID, if that is possible dont tell us it would been hard to put a screen up for every user logged in that says "We are aliasing cub with young. Do you agree yes/no?"

I have a feeling that any poll that might be interpreted as "please give your opinion on cub" would be a shitshow. It's on the default blacklist and a very touchy subject, to the point that there are people who do not want to be reminded of it at all.

regsmutt said:
I have a feeling that any poll that might be interpreted as "please give your opinion on cub" would be a shitshow. It's on the default blacklist and a very touchy subject, to the point that there are people who do not want to be reminded of it at all.

This too, I joked in the cub apocalypse topic about a 100 page discussion, we very well could have seen that were it announced across the site

We also no doubt would have seen numerous records and bans by extension

The forum users are 1% of 1% and we already saw multiple multi page topics about it with multiple records/bans

dba_afish said:
my dude, dragoness has been aliased away for over a decade, and, when it was active, it had only been added to a total of 100 posts. you don't know what you're talking about.

I just used it as a generic example of aliasing failing and even now, i cant actually search for female dragon pictures without swimming in pictures that has females and dragons in it but not an actual female dragon. Same can be said with tigress, lioness or any of the gendered species names the fandoms had.

regsmutt said:
I have a feeling that any poll that might be interpreted as "please give your opinion on cub" would be a shitshow. It's on the default blacklist and a very touchy subject, to the point that there are people who do not want to be reminded of it at all.

Yeah i can agree with that and from what i know about the booru system im not sure if it would be possible to run a poll that only asks people who have it removed from the blacklist.

lafcadio said:
Reply to this post with the exact search query you want to use, because you're almost certainly doing something wrong.

When that comment i replied to was made we still had loli and shota content so aliasing cub into young meant you still got furries and humans unless you used young_anthro which is longer. Its not much relevant now thought because that content disappeared thought its still the thing of replacing a 3 word tag with one that is longer and means the exact same thing.

strikerman said:
what kind of soul is in "cub"

Or style if you prefer that. Its like if you go into a japanese anime specialized imageboard, i expect them to stick to the theme and use kemonomimi, shota and others in the tags and not catgirl, young boy and similar generic aliases.

tester29 said:
When that comment i replied to was made we still had loli and shota content so aliasing cub into young meant you still got furries and humans unless you used young_anthro which is longer. Its not much relevant now thought because that content disappeared thought its still the thing of replacing a 3 word tag with one that is longer and means the exact same thing.

We still very much have content like that, so I'm not sure what your point is. Cub was thrown out specifically because its definition was inconsistent and it was underapplied. If you want to argue about aliasing tags in general, start a new topic as that is not relevant to this topic. This topic was over a year ago.

tester29 said:
I just used it as a generic example of aliasing failing and even now, i cant actually search for female dragon pictures without swimming in pictures that has females and dragons in it but not an actual female dragon. Same can be said with tigress, lioness or any of the gendered species names the fandoms had.

you literally said "anymore" like it was a tag that you had used in the past, when every gendered species tag has been aliased away since before your account was even created, but-- whatever.

also, we already talked about this: topic #56272.

donovan_dmc said:
We still very much have content like that, so I'm not sure what your point is.

I think this user was one of the handful of people who thought that loli/shota and cub were, like, mutually exclusive for some reason.

dba_afish said:
I think this user was one of the handful of people who thought that loli/shota and cub were, like, mutually exclusive for some reason.

I can "kinda" see it using shota/loli for human/humanoid content and cub only for anthros/feral but after that purge it makes even less sense now lol...

Still amazed how people can revive such old threads than making new ones, wonder if it would be possible to make that threads older than 6 months that aren't BUR/Etc related get auto locked or at least has as a warning that it's...old and it's better to make a new one.

notknow said:
...wonder if it would be possible to make that threads older than 6 months that aren't BUR/Etc related get auto locked or at least has as a warning that it's...old and it's better to make a new one.

It's arguable whether it's better or not; if you make a new one, everyone will just direct you to the old one.

Besides, I've been trawling through old threads for feature improvement ideas, and have 2 in particular that have been added to the site, and since most people won't check the version patch notes at the bottom of the page nor look at the updates thread, saying "this feature got added" in the place it's most relevant to do so is wise imo. Granted, that's very niche, but a relevant counterpoint nonetheless.

aacafah said:
Besides, I've been trawling through old threads for feature improvement ideas, and have 2 in particular that have been added to the site, and since most people won't check the version patch notes at the bottom of the page nor look at the updates thread, saying "this feature got added" in the place it's most relevant to do so is wise imo. Granted, that's very niche, but a relevant counterpoint nonetheless.

We used to utilise a feature request/bug report template that would allow developers/admins to edit the title and properly indicate whether some features were added or rejected.
However, because people are literal animals and can't even bother reading the pinned thread right in front of them and using the template properly, it became disused.

Nowadays, everything is on GitHub anyways so people (as in developers) can submit whatever feature they want for consideration and the status of the request will get updated over time.

notknow said:
I can "kinda" see it using shota/loli for human/humanoid content and cub only for anthros/feral but after that purge it makes even less sense now lol...

Still amazed how people can revive such old threads than making new ones, wonder if it would be possible to make that threads older than 6 months that aren't BUR/Etc related get auto locked or at least has as a warning that it's...old and it's better to make a new one.

The only reason i revived it was because i seen the changed, thought its recent, started typing in then seen its already a year old and....eh why not comment on it. Also because i replied to people, i grabbed more potential discussion then i think i would had with a new thread.
Also yes, i always used it as a separation. It technically means the same age group but separates the species as most places dont call fillies as shotas just because they are male, they use that exclusively for humans.

If any mod around thinks this post here is useless, i think they can close it fair and square considering its been revived.

thegreatwolfgang said:
We used to utilise a feature request/bug report template that would allow developers/admins to edit the title and properly indicate whether some features were added or rejected.
However, because people are literal animals and can't even bother reading the pinned thread right in front of them and using the template properly, it became disused.

Nowadays, everything is on GitHub anyways so people (as in developers) can submit whatever feature they want for consideration and the status of the request will get updated over time.

Only specific whitelisted contributors can add to that GitHub project, though any user can submit normal GitHub issues. Regardless, most users aren't going to go through the hassle of filing reports on another site (especially if they either don't have an account or don't want to file it on an account that employers would be looking at). We quarantine bug reports to its dedicated master thread, but we're always gonna have people requesting new features on here anyways. Besides, that doesn't do much for older feature request threads. Honestly, revamping the forums to be less... the way they are is a big wishlist item for me; using forms for certain boards, better filtering, etc. Maybe one day...

tester29 said:
...Also because i replied to people, i grabbed more potential discussion then i think i would had with a new thread...

...Do you think people are checking their email & rushing to e6 cuz something they wrote in 2 years ago had someone say "me too"? If you'd raised the same points, you'd have probably gotten the same response.

tester29 said:
Also yes, i always used it as a separation. It technically means the same age group but separates the species as most places dont call fillies as shotas just because they are male, they use that exclusively for humans.

That's not helping your case; to reiterate:

donovan_dmc said:
...Cub was thrown out specifically because its definition was inconsistent and it was underapplied...

aacafah said:

If you don't see the problem in that...

As dba afish said, this isn't news, but, since it seemingly wasn't expressed clearly enough before, I'll answer the question again.

Even if cub wasn't as fuzzily applied, we try to minimize redundant tags to prevent bloat (among other things). Gendered species tags are a better example.

Say you have a dragoness/female_dragon tag. Some users will only tag dragoness, & not tag female & dragon separately, which screws up results for people who are searching for female in general.

So, to fix this, you can create 2 implications; dragoness -> female & dragoness -> dragon. Problem solved? For this tag, yes. However, now people will come to rely on that expectation, & apply it elsewhere.

Well, we can add this to tigress, lioness, etc., right? Ok; what about horses? We're gonna need stallion & mare to maintain consistency. Plus, now other popular species without gender-specific names & people with a hankering for male tigers and herm dragons want in, and they're adding female_turtle male_tiger.

You could either nip this in the bud, or repeat this process as needed; I mean, we'll only do it for a few popular species, right? So we decide dragoness -> female_dragon & female_dragon -> female & female_dragon -> dragon, and repeat this for every gender & the most popular species. We've went from 2 tags & no implications to 7 (8 w/ intersex) combined tags, with each requiring 2 implications to their separated forms to maintain searchability, times however many species we decided. But hey, at least we're done, right?

Wrong; the more the popular species follow this trend, the more people will apply it to less popular species. So now, people are adding andromorph_klefki & expecting things to work out; now no one's getting what they want.

So now, we need to make proper implications for every single species & every single gender to make sure the vast majority of posts are actually consistently tagged for the 2 biggest things people are searching for. Also, remember that some specifically formatted/commonly misspelled tags are already aliased, so we now need extra implications for those.

But once we do that, that should be the end of it, right??? No.

Because, now that the 2 most well-tagged categories (gender & species) behave like this, people will naturally assume most other tags work like this (hi again cub). At this point, we have completely lost control of this circus; the server has to decompose a tangled rat's nest of aliases & implications for every upload & every search, some posts become bloated with tags, and despite this people can't search consistently w/o having to add the separated & combined forms of some tags (which was, you know, the whole point of this exercise), & new users just follow what they see, exacerbating the issue.

Or... you could just type dragon female & be reasonably confident you'll get every post with a dragon & a female on it, & be done with it.

the only possible way we could have <adjective>_<species> would be to cut it off way at the "top-level" taxa so, *_reptile, *_mammal, *_arthropod, *_mollusk, etc. although, you'd really still need to figure out what "top level" is for every character on the site, and even then, what with hybrid and animal_humanoid and whatnot that still gets pretty fucky. form and gender are a little bit more concrete, since, with how we define them, every character should have exactly one form and gender at any single moment, while with species a character could feasibly be any number from any number of orders, kingdoms, or even biological bases.

and even if we did manage to find really good concrete lines to draw on what species tags get the "special noun" treatment, and decided what exact roles they're allowed to fill, I wouldn't be surprised if this still lead to a mouse-cookie problem that we'd need to keep dealing with.

On top of that, we'd still have people who tag like we do now & only use the decomposed forms; unless the combined forms are consistent & predictable, these users' numbers would increase, further crushing the utility of this option.

aacafah said:
...Do you think people are checking their email & rushing to e6 cuz something they wrote in 2 years ago had someone say "me too"? If you'd raised the same points, you'd have probably gotten the same response.
That's not helping your case; to reiterate:

If you don't see the problem in that...

As dba afish said, this isn't news, but, since it seemingly wasn't expressed clearly enough before, I'll answer the question again.

Even if cub wasn't as fuzzily applied, we try to minimize redundant tags to prevent bloat (among other things). Gendered species tags are a better example.

Say you have a dragoness/female_dragon tag. Some users will only tag dragoness, & not tag female & dragon separately, which screws up results for people who are searching for female in general.

So, to fix this, you can create 2 implications; dragoness -> female & dragoness -> dragon. Problem solved? For this tag, yes. However, now people will come to rely on that expectation, & apply it elsewhere.

Well, we can add this to tigress, lioness, etc., right? Ok; what about horses? We're gonna need stallion & mare to maintain consistency. Plus, now other popular species without gender-specific names & people with a hankering for male tigers and herm dragons want in, and they're adding female_turtle male_tiger.

You could either nip this in the bud, or repeat this process as needed; I mean, we'll only do it for a few popular species, right? So we decide dragoness -> female_dragon & female_dragon -> female & female_dragon -> dragon, and repeat this for every gender & the most popular species. We've went from 2 tags & no implications to 7 (8 w/ intersex) combined tags, with each requiring 2 implications to their separated forms to maintain searchability, times however many species we decided. But hey, at least we're done, right?

Wrong; the more the popular species follow this trend, the more people will apply it to less popular species. So now, people are adding andromorph_klefki & expecting things to work out; now no one's getting what they want.

So now, we need to make proper implications for every single species & every single gender to make sure the vast majority of posts are actually consistently tagged for the 2 biggest things people are searching for. Also, remember that some specifically formatted/commonly misspelled tags are already aliased, so we now need extra implications for those.

But once we do that, that should be the end of it, right??? No.

Because, now that the 2 most well-tagged categories (gender & species) behave like this, people will naturally assume most other tags work like this (hi again cub). At this point, we have completely lost control of this circus; the server has to decompose a tangled rat's nest of aliases & implications for every upload & every search, some posts become bloated with tags, and despite this people can't search consistently w/o having to add the separated & combined forms of some tags (which was, you know, the whole point of this exercise), & new users just follow what they see, exacerbating the issue.

Or... you could just type dragon female & be reasonably confident you'll get every post with a dragon & a female on it, & be done with it.

Hell, we have grouped tag searches now; you could do ~( dragon female -male ) ~( dragon female solo ) , and get specific with other search terms, etc. It's a worthy trade-off to have many consistent general tags over fewer specific tags; the only way to have both is to have half the community racing the other half to clean up their mess before they make another one.

Well if you check back here, my first reply was a complaint on how using the blacklist along with typing in young is longer than just typing in cub. I dont think this point would gather much attention nor the next one when i said it should been announced if actual serious user input is needed.

As for the tagging system i had a thread with this exact issue and suggested the ehentai approach where every species tag is gendered by default.

You have a picture with a female dragon and a male lion.

In the upload form theres a line saying "Species" and a plus symbol for adding more species. Click on the species one, type in dragoness and it immideatly prompts you if the tag exists or not and prompts you to choose gender for it. When finished you click the plus icon, type in lion and as before immideatly gender it then add the rest of the tags. The end result looks like this:

Female: Dragoness, Dragon(the system automatically adds this because of dragoness), Anal, Happy expression
Male: Lion, Yellow fur, blind

Everything is interconnected and refers to the other while automatically filling in the necesseary parts but this is as i heard not possible on the booru system.

tester29 said:
... i had a thread with this exact issue...

I know. I linked it.

...As dba afish said, this isn't news...

You had a hanging question...

...I just found it strange that the gendered tags were removed but no function to cover them exist. Why were they removed initially?...

...so I attempted to more thoroughly answer it.

tester29 said:
...but this is as i heard not possible on the booru system.

Not without rewriting half the relevant backend & frontend logic & retagging over 5 million posts, no.

tester29 said:
Well if you check back here, my first reply was a complaint on how using the blacklist along with typing in young is longer than just typing in cub. I dont think this point would gather much attention nor the next one when i said it should been announced if actual serious user input is needed.

As for the tagging system i had a thread with this exact issue and suggested the ehentai approach where every species tag is gendered by default.

You have a picture with a female dragon and a male lion.

In the upload form theres a line saying "Species" and a plus symbol for adding more species. Click on the species one, type in dragoness and it immideatly prompts you if the tag exists or not and prompts you to choose gender for it. When finished you click the plus icon, type in lion and as before immideatly gender it then add the rest of the tags. The end result looks like this:

Female: Dragoness, Dragon(the system automatically adds this because of dragoness), Anal, Happy expression
Male: Lion, Yellow fur, blind

Everything is interconnected and refers to the other while automatically filling in the necesseary parts but this is as i heard not possible on the booru system.

creating a nested tagging relationship system like this would probably be totally unworkable in practice, certainly
more of a mess than any other possible method. not only would it require a total rework of the tagging system in general and require that we figure out which tags can be the child of which other tags, it'd also mean that we'd need to figure out a way for users to actually, y'know, tag and search the stuff, because tagging a character wouldn't just be something like:

solo female anthro dragon red_scales nude big_breasts non-mammal_breasts vulva

etc.

it'd be more like:

solo { 
   female { 
      anthro {
         dragon {
            red_body {
               red_scales
            }
            nude {}
            vulva {}
            breasts {
               non-mammal_breasts
               big_breasts
            }
         }
      }
   }
}

and even that breaks down the instant you do anything beyond single static images of non-hybrid characters. animations, comics, _TF_? it just wouldn't work.

I'm in the middle of tagging projects that a site has... very bad Booru imports from a Booru with some of the worst tagging I've ever seen. It's actually progressing, amazingly.

This thought has been in my head for months, so let me just put it all down here so I can get feedback & link people to the next time this comes up.

dba_afish said:
...creating a nested tagging relationship system like this...

...tagging a character wouldn't just be something like:

solo female anthro dragon red_scales nude big_breasts non-mammal_breasts vulva

etc.

it'd be more like:

solo { 
   female { 
      anthro {
         dragon {
            red_body {
               red_scales
            }
            nude {}
            vulva {}
            breasts {
               non-mammal_breasts
               big_breasts
            }
         }
      }
   }
}

...

Not that I think it's plausible, but this isn't really nested tagging, it's character-based tagging; in storage, it'd probably look more like

arty_mcartface_(artist) duo hi_res ~1~ female anthro dragon... ~2~ male anthro tiger...

You'd put non-character related tags in front (e.g. artist tags, meta tags, character count tags, etc.), then, for each character, adding a delimiter to denote the next section of tags is for the same character followed by the tags applying to that character. Doing it this way would mean searching irrespective of which character the tags are on would likely be wholly unchanged. The syntax for including them would just be the usage of some restricted character (let's just use : for example) + a number to uniquely id that character in this search & prefacing the tags related to that character with that (e.g. :1:female :1:dragon :2:male :2:tiger). Parsing the character-centric tags & actually breaking down the search string to search for all those tags on the same character is where the trouble starts (+ the current assumption tags are always in alphabetical order & a million other little things). Parsing is tedious but doable (unless you want to use character identifiers inside a grouped search & have the unique character id to refer to the same character, in which case, good luck - you'd probably have to rewrite the entire feature; take it from the guy who made it), & actually formatting that into a request the search service can handle is... ambitious. You could probably use an intervals query with either a regexp or a range rule, but that would probably tank performance, although that could be offset by restricting the limit of 40 tags for these searches or highly restricting either the # of characters and/or character-specific tags these searches can use, but again, a lot of ifs there, & a lot of performance profiling & tuning.

dba_afish said:
...and even that breaks down the instant you do anything beyond single static images of non-hybrid characters. animations, comics, _TF_? it just wouldn't work.

I think hybrid characters & single image TF would work with the outlined approach, but animations, comics, & sketches would stumble by having tags that are accurate in 1 panel be inaccurate elsewhere, + image tags like monochrome, colorful, looking_at_viewer, etc. could apply to a single panel/shot instead of the work as a whole (unless this is expanded to mark up panels & shots, which, ew, no); however, would it be any worse than it does now?

The fact that this could be done while not interfering w/ preexisting posts (and, depending on implementation, new posts tagged like the old ones) makes this viable.

In short, it is possible, but will likely never happen due to potentially contentious UX changes, performance considerations, excessive development time, & lack of confidence that the value derived is worth the cost to add.
Thank you for coming to my TED talk.

aacafah said:
-snip-

part of the problem I was having was trying to conceptualize how we'd associate tags to a parent. at first I was thinking about having "dummy" tags for unnamed characters, but even that would have problems. having it work as more of a "each tag can be associated with (an) index values" sorta deal would probably work more conceptually. it would still have other problems, most notably getting people to actually use it.

it feels like it'd have to be a system built, like, next to tagging, rather than being directly integrated into the tagging system. so, each post gets tagged and then each character index can be associated with any of those tags.

I feel like this system actually could have some nominal utility if we made it work.

        ...
ʚ(ϵ ⁰–⁰)϶

honestly, now I'm just thinking about how a system like this could maybe finally work out the inside and outside tags. we could start associating the location tags with individual indexes and maybe have an index 0 that was always for the "camera" or maybe a PoV character.

almost certainly still not worth the effort, though.

dba_afish said:
part of the problem I was having was trying to conceptualize how we'd associate tags to a parent. at first I was thinking about having "dummy" tags for unnamed characters, but even that would have problems. having it work as more of a "each tag can be associated with (an) index values" sorta deal would probably work more conceptually. it would still have other problems, most notably getting people to actually use it.

it feels like it'd have to be a system built, like, next to tagging, rather than being directly integrated into the tagging system. so, each post gets tagged and then each character index can be associated with any of those tags.

I feel like this system actually could have some nominal utility if we made it work.

        ...
ʚ(ϵ ⁰–⁰)϶

honestly, now I'm just thinking about how a system like this could maybe finally work out the inside and outside tags. we could start associating the location tags with individual indexes and maybe have an index 0 that was always for the "camera" or maybe a PoV character.

almost certainly still not worth the effort, though.

There's some weirdness with classifier systems like this. In some hypothetical higher-dimensional space, the shape for a 0 or a 2 or a 5 is nearer say, OQ* or 7Z or S$, respectively. There's some non-trivial examples like unknown characters that works like this. A neat thing is to try image search on an unknown character on Yandex image search. Even if you don't find the right character, you'll get tons with similar designs!

It's funny because doing this by hand is one of the very things that the classifiers are far better at than humans when automating. You get 'fuzzy' matches, that can then be later fixed once someone fills in the gaps by manually adding the tag to similar results.

dba_afish said:
...it feels like it'd have to be a system built, like, next to tagging, rather than being directly integrated into the tagging system. so, each post gets tagged and then each character index can be associated with any of those tags....

That would be less contentious, but that'd potentially make searches more taxing & require a massive modification to the db; instead of simply scanning the tag_string field for the tag appearing anywhere (as we do now & would in the proposal) or scanning the tag_string field for tags with the same id in the same range (as in the proposal), that'd require the search system to index into another table (more likely) or scan a separate new field on post records (less likely). Regardless of which way you choose to skin this cat, it's gonna be a mess.

dba_afish said:
we could start associating the location tags with individual indexes and maybe have an index 0 that was always for the "camera" or maybe a PoV character.

I'd say that's a good example, although I'd say location tags that refer to point-of-view of the image as a whole would be classified as "non-character related tags", & could just be placed in front of the character-related tags like artist tags & meta tags in the example, with location tags referring to the location of a character being placed in the appropriate ranges. To tweak the prior example, arty_mcartface_(artist) duo hi_res outside ~1~ female anthro dragon inside... ~2~ male anthro tiger inside... would describe a hi res image with 2 characters in it, drawn by arty_mcartface_(artist), with a pov from outside a structure, looking in on 2 characters; an anthro female dragon & an anthro male tiger. The only exception I could think of is first person view, so that label should belong to the image as a whole & pov character specifically. I think allowing people to search for non-character-specific tags has little utility (almost all tags that apply to the image as a whole only apply to the image as a whole) & adds complications, so I'm not sure about allowing those searches, but idk.

alphamule said:
There's some weirdness with classifier systems like this. In some hypothetical higher-dimensional space, the shape for a 0 or a 2 or a 5 is nearer say, OQ* or 7Z or S$, respectively.

...what? If you're talking about higher-dimensional vectors being used to find relations (like with machine learning), I don't see the relation to this. If not, I'd be curious to learn.

alphamule said:
It's funny because doing this by hand is one of the very things that the classifiers are far better at than humans when automating. You get 'fuzzy' matches, that can then be later fixed once someone fills in the gaps by manually adding the tag to similar results.

Assuming I'm correct about what you're discussing, that's not unique to relational data; e6 uses OpenSearch/ElasticSearch, which does have the capability to calculate relevance scores based on how many terms are matched by the search & more; we just override that in favor of sorting by id (or vote score, date created, etc.) for performance & possibly other reasons I'm unaware of.

aacafah said:
-snip-

Seems like you are attempting to make something more analogous to https://en.wikipedia.org/wiki/Resource_Description_Framework . I have worked on a tagging system (albeit, proprietary and internal to my employer) that used this format, with a triplestore backing it. From experience, it's complicated for people to search such stores, not to mention conceptualizing it. However, it's powerful and the ability to search semantics really allows you to find exactly what you are looking for.

If you want an example of something actually using RDF today, take a look at Wikidata. It's run by the Wikimedia foundation and applies RDF metadata to many pieces of info within Wikipedia, using their own ontology. You can search using SPARQL (the standard query language for RDF, equivalent to SQL for structured data).

IMO, it's better to use standard tech rather than trying to reinvent it. Semantic Web tech would be my starting point for any effort like this. Unfortunately, it requires a lot of mental and technical overhead that makes it a non-starter in many cases.