I've noticed that there are some characters that have more than one character tag, especially when it comes to 3D art of them that use a specific popular model.

For example, loona_(aeridiccore) and loona_(carbiid3) both depict the character loona_(helluva_boss) and yet do not implicate the main character tag, as if they were different characters completely. Another example of this is seen where krystal_(dogzeela) and warfare_krystal both reference the character of krystal, yet there are also no corresponding tag implications. There are many more examples throughout the site, but these two characters are particularly notable.

Given this situation, there are three options that are immediately apparent:
1. Leave this informal system as-is
2. Find these "model tags" and make them implicate their respective characters
3. Remove the "model tags" entirely, as they are not separate characters in and of themselves

What course should be taken? Or are there any other compromises that might be able to be made?

I say option 3. They're not characters so shouldn't be tagged as characters; they're basically designs of characters and species, which we don't tag (at best we tag different canon forms of characters, e.g. link_(wolf_form), but we don't tag link_(elranno) for Elranno's distinctive design of Wolf Link or link_(uzucake) for Uzucake's distinctive design of Link as a wolf). It creates a mess of the character tags on 3D posts, as every character would always have two associated character tags; one for the character themself, and one for whoever made the model. Just by virtue of being a 3D image instead of 2D.

Every species would also be tagged for every individual model that gets made, and it allows people to get around hybrid tagging restrictions by making a 3D image and tagging it as a model/character, despite not being a character (we don't tag vaporunny, but we can have as many vaporunny_(modeller) tags that people can make?). We don't tag generic hybrids, no matter how identifiable the hybrid design may be, unless it becomes distinct enough to be its own species or character, and I don't see why that should be different for 3D images. We also don't tag who modeled individual elements like trees or penises (and there are very commonly used penis models that are easily identifiable), so I don't see why character models should be treated specially.

At best I can see a case for tagging a modeller like petruz or warframe, but I don't see any reason to have unique character and species tags for each individual model of theirs. character_name+modeler or species_name+modeler would be enough to search for the given modeler's species or character model with a low amount of false positives.

While I'm actually pro tagging models, I can definitely see that Watsit’s argumentation is correct.
Nevertheless we need some tag that enables users to either search or blacklist models (You know which FNAF models I’m talking about), so tagging the modeler sounds like the best solution. Would also help against the tag bloat from all the different models.

To atomicBlaze21’s suggestion of making the models imply their respective characters: That wouldn’t work. It happens often that one model is used for different characters. For example in the case of disembowell’s infamous FNAF models the model of Toy Bonnie is also used for Bonnie, Shadow Bonnie, Springtrap and Spring Bonnie.

I also have a personal interest in this discussion, because a while ago I tagged almost all of disembowell’s FNAF models and it later turned out that there was a mistake. We thought these models belonged to psychjohn2 and tagged them as such. As of now the models still have the incorrect tags and I would really like to fix that.

Another question that I brought up in the older forum post about models (https://e621.net/forum_topics/40657 ) and maybe someone on the staff could answer: Can modelers make a takedown request for models they created? I mean, they made the model, it is their work.

I would say 3. They are all the same character, just drawn/modelled differently.

The common argument is for users who dislike the art style to blacklist the artist entirely, but since this is not feasible for pre-made models being used by multiple artists, a suitable solution should be in order first before the removal of this informal system.

Tagging models is fine. I would say the larger problem was users lacking restraint--a common problem--trying to track down literally every model ever used for a given character for each piece of 3D art that found its way onto e621. Seemed very unnecessary. If something is so obscure that you have to ask, then it might not be much of an issue for blacklisting, but then the opposite consideration still remains. Searchability.

Ideally, everything recognizable is searchable, including specific models that users may greatly prefer over others. For example, I like the detailed Hornet model used in post #4295543, but I think the other simple, more lore-accurate Hornet models are unappealing. Should e621 be allowed to serve my interests? Right now, we don't have too much 3D art of Hornet so my search burden is small, but what if I preferred a specific model of a FNAF character or charizard? I'd be shit out of luck, and as it happens there is a charizard model I kind of like. Guess I'm screwed, right?

"We don't do that" is a terrible, shortsighted argument. We would never do anything new if that held any water. Cub was a tag last week with 145K posts and had been a tag for more than 15 years, but as it's turned out "we don't do that" anymore. Even though the cub wiki, written by staff, clearly laid out how to use the tag, staff created contradictory tag locks, so sometimes "what we do" isn't even "what we do."

Further, anyone can edit tags, we probably have 10s of thousands of active users, and charitably we have 50-100 users who fix tags according to "what we do." Perfectly maintaining "what we do" is impossible. New wrinkles creep into the system every day that go unchecked for months, years, or more than a decade. These wrinkles that no one agreed to sometimes persist so long that they become embedded in e621's DNA, causing many users to believe that actually "we do that." What's more, old users leave who would have remembered "what we do" as decided 5-10 years ago, and new users are unable to divine these old decisions, many of which were only informally laid out in our poorly searchable forums. "What we do" is ephemeral, true one moment and not the next.