The bulk update request #1218 is active.

remove implication calico_cat (7242) -> piebald (4205)
create implication calico_cat_humanoid (44) -> cat_humanoid (31496)
create implication norwegian_forest_cat_humanoid (1) -> cat_humanoid (31496)
create implication bobcat_humanoid (7) -> lynx_humanoid (22)

Reason: Implicating a species tag to a physical characteristic is not done elsewhere in e621 as far as I know, and it could lead to conflict with the TWYS principle.

Specific breeds of domestic cat imply domestic_cat, so logically specific breeds of cat humanoid should imply cat_humanoid. The same argument applies to species in the feline taxon as well.

EDIT: The bulk update request #1218 (forum #315278) has been approved by @gattonero2001.

gattonero2001 said:
Specific breeds of domestic cat imply domestic_cat, so logically specific breeds of cat humanoid should imply cat_humanoid.

They should also imply the breed of cat.

create implication calico_cat_humanoid -> calico_cat
create implication norwegian_forest_cat_humanoid -> norwegian_forest_cat
create implication bobcat_humanoid -> bobcat

Species of humanoids do not imply the general tag as a rule. In my opinion, the *_humanoid system and the general one should be kept entirely separate.

gattonero2001 said:
Species of humanoids do not imply the general tag as a rule. In my opinion, the *_humanoid system and the general one should be kept entirely separate.

feline_humanoid implicates feline. So if someone searches feline, they get all felines, anthro, feral, and humanoid. But if someone searches bobcat, they'll only get some bobcats, anthro and feral only and no humanoids. That doesn't seem logical or helpful.

watsit said:
feline_humanoid implicates feline. So if someone searches feline, they get all felines, anthro, feral, and humanoid. But if someone searches bobcat, they'll only get some bobcats, anthro and feral only and no humanoids. That doesn't seem logical or helpful.

If it were up to me, feline would return only anthros and ferals, just like domestic_cat and felis already do. There is no point in mixing up the results with the higher taxons when we practically almost have two separate evolutionary trees of tags.

watsit said:
feline_humanoid implicates feline. So if someone searches feline, they get all felines, anthro, feral, and humanoid. But if someone searches bobcat, they'll only get some bobcats, anthro and feral only and no humanoids. That doesn't seem logical or helpful.

This also seems to be the case with dragon_humanoid and, well, most other species tags I've seen. I don't see why not keep it that way... you can always search with anthro feral or -humanoid if that's not what you're looking for.

scaliespe said:
This also seems to be the case with dragon_humanoid and, well, most other species tags I've seen. I don't see why not keep it that way... you can always search with anthro feral or -humanoid if that's not what you're looking for.

Most other species tags I've seen do not follow dragon_humanoid's example. Having a regular species search return anthros, ferals and humanoids at the same time is worse than returning only anthros and ferals, because it removes granularity for no reason.

Let's look at two hypothetical situations:

• If person A searches the tag domestic_cat, the site will return only anthros and ferals, which is what A wants.

• If person B wanted to see cat humanoids as well, B could make a separate search OR use the tilde operator, a simple feature that is described in the "Help" link at the top of every page. ~domestic_cat ~cat_humanoid

• If person A searches the tag dragon, the site will return anthros, ferals and humanoids, which is not what A wants.

• If A doesn't want to see dragon humanoids, using -humanoid would solve that issue, but it would also remove posts that have both anthros and/or ferals AND humanoids, that A might be interested in.

• If A searches dragon AND anthro and/or feral, there is no guarantee that the "anthro" and/or "feral" in a given post refers to a dragon. The search might return a dragon humanoid and a feral bird, for example, which is definitely not what A wants.

The first situation has a very simple and immediate fix for B, everyone is happy.

The second situation has no satisfactory fix for A.

Ideally, there would be *_anthro and *_feral tags for every species as well, but that ship has sailed. It could have been done early in the site's history, but now there is too much inertia from the current system.

However, that is not the case for *_humanoid. The system is already in place and there is no reason to mix humanoids with anthros/ferals.

gattonero2001 said:
Most other species tags I've seen do not follow dragon_humanoid's example. Having a regular species search return anthros, ferals and humanoids at the same time is worse than returning only anthros and ferals, because it removes granularity for no reason.

Let's look at two hypothetical situations:

• If person A searches the tag domestic_cat, the site will return only anthros and ferals, which is what A wants.

• If person B wanted to see cat humanoids as well, B could make a separate search OR use the tilde operator, a simple feature that is described in the "Help" link at the top of every page. ~domestic_cat ~cat_humanoid

• If person A searches the tag dragon, the site will return anthros, ferals and humanoids, which is not what A wants.

• If A doesn't want to see dragon humanoids, using -humanoid would solve that issue, but it would also remove posts that have both anthros and/or ferals AND humanoids, that A might be interested in.

• If A searches dragon AND anthro and/or feral, there is no guarantee that the "anthro" and/or "feral" in a given post refers to a dragon. The search might return a dragon humanoid and a feral bird, for example, which is definitely not what A wants.

The first situation has a very simple and immediate fix for B, everyone is happy.

The second situation has no satisfactory fix for A.

Ideally, there would be *_anthro and *_feral tags for every species as well, but that ship has sailed. It could have been done early in the site's history, but now there is too much inertia from the current system.

However, that is not the case for *_humanoid. The system is already in place and there is no reason to mix humanoids with anthros/ferals.

As you’ve noticed, the same issue exists with ferals and anthros, so only fixing it regarding humanoids is a bit inconsistent. And it’s actually a much worse problem with ferals, since a feral can be something like a bird or insect sitting somewhere in the foreground, which functions more as a decoration for the scene than an actual character. -feral will probably remove otherwise relevant results more often than the other body type tags.

Not to mention that there is a definite gray area between anthro and humanoid. Is undyne a fish anthro or a fish humanoid? You can find posts going either way.

Another issue is with hybrids, chimeras, and split forms. Even in your first scenario, person A will get cat hybrids and other mixed creatures, which may not even resemble cats very closely, which is probably not what they want; and searching -hybrid -chimera -split_form may remove results that they do want.

Then what about taurs and lamias? Currently, taur follows the same pattern as humanoid - we have tags like dragon_taur and cat_taur that also implicate dragon and domestic_cat, respectively. Lamias don’t even get tags like dragon_lamia—they are just tagged with lamia and the regular species tag separately, which further complicates the issue.

Whatever is done, ideally, it should be consistent across these various body types. Either follow the species_bodytype formula for all body types, or for none of them. Sure, fixing anthro and feral at this point would be a massive effort, but so would removing all the regular species tags on humanoids. There are nearly 100,000 results for animal_humanoid, most of which will have to be fixed if we are to remove regular species tags from humanoids. Then there are still many more such as animal_heads and plant_humanoids that usually get the species tag, so the total number of posts that need to be fixed is probably much greater than 100,000.

The only true solution I can imagine would be to implement a tag grouping system, whereby each group of tags relates to an individual character, and users could search by grouping tags together, ie. [ anthro dragon male ] [ humanoid cat female ]. This would require a major change to the tagging system, however; and given the rate at which simple aliases and implications get implemented, it’s not likely to happen any time soon. As for our current predicament, if we aren’t going to go through some major change to enforce consistency between these tags (either adding *_anthro *_feral etc. tags or removing *_humanoid *_taur), then the only thing that really makes sense is to continue doing what we’ve already been doing, as imperfect as it is.

scaliespe said:
As you’ve noticed, the same issue exists with ferals and anthros, so only fixing it regarding humanoids is a bit inconsistent. (…)

Another issue is with hybrids, chimeras, and split forms. (…)

Then what about taurs and lamias? (…)

Whatever is done, ideally, it should be consistent across these various body types. (…)

We can't let "perfect" be the enemy of "good enough". Sacrificing functionality to stay "consistent" is absolutely not a worthy tradeoff. Humanoids are fundamentally different from anthros, ferals and other forms because a substantial proportion of them are essentially anime girls with animal ears and therefore tangential to the interests of most e621 users.

Implicating cat humanoids (and/or their respective breeds) to the base tags would be absurdly detrimental to the search results, more so than possibly any other humanoid -> base implication, since anime catgirls are vastly overrepresented among humanoids.

• dragon_humanoid: less than 4000 posts

• cat_humanoid: over 22000 posts

That is probably the reason why the lower taxon implications were avoided when the higher taxon implications were created. Which doesn't mean that the higher taxon implications themselves were a good idea.

Not to mention that there is a definite gray area between anthro and humanoid. Is undyne a fish anthro or a fish humanoid? You can find posts going either way.

Characters may change form and/or species (and/or gender, among other characteristics) in different depictions. The original game sprite is an anthro, but when Undyne is drawn as a regular anime girl with blue skin and finned ears then that would be a humanoid.

The only true solution I can imagine would be to implement a tag grouping system, whereby each group of tags relates to an individual character, and users could search by grouping tags together (…)

A cat can dream… Yes, that would be wonderful.

I do not quite agree with the conclusion of your paragraph, though. As I already said, sacrificing functionality to stay "consistent" is not worth it. It makes sense to treat humanoids as a distinct group from other forms and that is specially critical regarding cats.

To be clear, I would not be pushing for the complete separation of "species humanoids" and "general species forms" if a partial separation did not already exist. It just feels like missed potential. I hope until the end of 2021 we can see the removal of the higher taxon implications, at least.

The tag grouping idea is probably extremely difficult to implement, but if there were a lot more contributors to the code base I believe it could be done in a few years. Sadly, I find myself to be profoundly lacking in that department…

gattonero2001 said:
We can't let "perfect" be the enemy of "good enough". Sacrificing functionality to stay "consistent" is absolutely not a worthy tradeoff. Humanoids are fundamentally different from anthros, ferals and other forms because a substantial proportion of them are essentially anime girls with animal ears and therefore tangential to the interests of most e621 users.

Implicating cat humanoids (and/or their respective breeds) to the base tags would be absurdly detrimental to the search results, more so than possibly any other humanoid -> base implication, since anime catgirls are vastly overrepresented among humanoids.

• dragon_humanoid: less than 4000 posts

• cat_humanoid: over 22000 posts

That is probably the reason why the lower taxon implications were avoided when the higher taxon implications were created. Which doesn't mean that the higher taxon implications themselves were a good idea.

I can sympathize with this, truly, but adding -cat_humanoid is still going to give you 99.9% of relevant results. See the point about Undyne below for why I still think this may be preferable.

Characters may change form and/or species (and/or gender, among other characteristics) in different depictions. The original game sprite is an anthro, but when Undyne is drawn as a regular anime girl with blue skin and finned ears then that would be a humanoid.

Right, but that’s my point. Even in her original depiction, she lacks a tail. While a tail is not necessary for a character to be considered anthro, that detail (along with humanoid arms rather than fins or finned arms) puts her right inside the gray area between anthro and humanoid. Canonically, she doesn’t have a nose, yet some people draw her with a nose, and that may be the only difference I can find between an anthro Undyne and a humanoid Undyne. Given that there are thousands of images of her on here, this is a case that’s going to come up very frequently. So here’s the issue: if you’re looking for fish anthro but not fish humanoid, fish -fish_humanoid is still going to get a lot of results of Undyne. Maybe that’s what you want, or maybe it isn’t. Is the lack of a nose really enough to make a difference?

Sure, you could just -Undyne, but Undyne is just one example. This happens all the time with dragons, for instance. Having a humanoid face on an otherwise anthro dragon seems to be enough to warrant the use of the dragon_humanoid tag. Examples:
post #2789785
post #2746703
post #2850747

If you’re looking for dragons, are these relevant or irrelevant?

A cat can dream… Yes, that would be wonderful.

I do not quite agree with the conclusion of your paragraph, though. As I already said, sacrificing functionality to stay "consistent" is not worth it. It makes sense to treat humanoids as a distinct group from other forms and that is specially critical regarding cats.

To be clear, I would not be pushing for the complete separation of "species humanoids" and "general species forms" if a partial separation did not already exist. It just feels like missed potential. I hope until the end of 2021 we can see the removal of the higher taxon implications, at least.

The tag grouping idea is probably extremely difficult to implement, but if there were a lot more contributors to the code base I believe it could be done in a few years. Sadly, I find myself to be profoundly lacking in that department…

It may be for the best overall, but I still have reservations about it, particularly because edge cases are so common. But there still remains that issue of a) de-implicating all the humanoids, and b) cleaning up somewhere in the ballpark of 100,000 humanoid posts that also contain the general species tags. That second point would be a huge effort… probably even less of an effort than programming in tag grouping.

gattonero2001 said:
Implicating a species tag to a physical characteristic is not done elsewhere in e621 as far as I know...

It happens, but the implied characteristics are told on the tag itself, examples:

• feathered_dragon, feathered_snake and feathered_dinosaur implys feathers
• furred_dragon and furred_snake implys fur
• winged_unicorn implys horn and wings
• horned_humanoid implys horn

gattonero2001 said:
...as far as I know, and it could lead to conflict with the TWYS principle.

Could you show us examples that could cause conflicts with TWYS or even hypothetical scenarios? I've been looking for some, but couldn't.

The definition of a calico_cat in the Wiki states that a calico cat has a piebald coat.
Wikipedia of the Calico Cat states that "a calico cat is a domestic cat of any breed with a tri-color coat."

So I'd say that the imply calico_cat to piebald is okay. I don't see reasons to remove it.

scaliespe said:
If you’re looking for dragons, are these relevant or irrelevant?

All of your examples seem to be correct uses of the dragon_humanoid tag as far as I am concerned.

It may be for the best overall, but I still have reservations about it, particularly because edge cases are so common. But there still remains that issue of a) de-implicating all the humanoids, and b) cleaning up somewhere in the ballpark of 100,000 humanoid posts that also contain the general species tags. That second point would be a huge effort… probably even less of an effort than programming in tag grouping.

Regarding the implication removals, I have already written a BUR that does exactly that. It didn't have enough support to pass through the first time, so I am waiting for a change in public opinion to open a formal request again. However, the script can be found here.

Regarding the cleanup, I admit that it would not be easy, but it might be easier than you think. Using bitWolfy's TagMe!, several users can contribute to the project in a simple and standard way. I volunteer with enthusiasm.

sieghelm_lockayer said:
The definition of a calico_cat in the Wiki states that a calico cat has a piebald coat.
Wikipedia of the Calico Cat states that "a calico cat is a domestic cat of any breed with a tri-color coat."

So I'd say that the imply calico_cat to piebald is okay. I don't see reasons to remove it.

The wiki page for piebald says the following:

Piebaldism is an animal skin pattern, in which the animal has an unpigmented background of fur, hair, feathers, scales, etc that is covered with pigmented spots.

If a calico is any cat with a tricolor coat, that does not require an unpigmented background. It could theoretically be the opposite, and while that might be an edge case it is enough to invalidate such a broad implication.

"Calico" refers only to a color pattern on the fur, from colorful printed Calico fabric, not to a cat breed or any reference to any other traits, such as its eyes. - Wikipedia

So it could be argued that the "calico" part could be a implied characteristic for piebald as the tags I aforementioned. The calico cat is also know as tricolor cat and in some languages this characteristic is even more implied. In Japanese they are know as 三毛猫 (Mikeneko) 'triple fur cat'. In Portuguese they are also know as Gato de Chita, Gato Malhado (Coated Cat) or Gato Tricolor (Tricolor Cat). The "Malhado" part can even be translated as piebald.

"Malhado" would have a closer meaning to "meshed", in my opinion. But that is beside the point, as popular nomenclature does not describe the physical characteristics of an animal with high precision.

gattonero2001 said:
All of your examples seem to be correct uses of the dragon_humanoid tag as far as I am concerned.

Sure, I’m not suggesting that they’re incorrect. I’m suggesting that they’re so close to being anthro that pretty much anyone who is searching the dragon tag will probably be okay with those results. It’s far from the cat-eared anime girls you’re taking issue with. I assume the same would probably be the case with cats - there are probably some who are covered in fur and have all cat features but with a humanoid face, and are so tagged as humanoid. This is what I mean by edge cases. As far as I’m concerned, the character is still like 90% dragon or cat or whatever, and should still come up under the species tag. Perhaps what we need is some intermediary between anthro and humanoid, like what semi-anthro is to ferals.

scaliespe said:
(…)

In my opinion, there is a fundamental difference between those "edge cases" and anthros that is not present between the former and the average humanoid. Sure, a special tag (in addition to *_humanoid, but not using the general tag) to make them stand out from the rest might be interesting.

However, that issue is unrelated to (and out of the scope of) the above BUR in this topic. It might be better to continue the discussion in a new one.

The bulk update request #1218 (forum #315278) has been approved by @gattonero2001.