Bethesda Games have accumulated a lot of fragmented modding communities, and various body mods, especially for female player characters.

And, there are varying degrees of compatibility between these accumulations of work:

1. Compatibility across games. Some tools (like NifSkope) offer some forms of this.
2. Compatibility across gender. Again, this is an issue with tools.
3. Compatible body shapes (this is what makes clothing [that's not body-replacer clothing] work without clipping -- also I'm ignoring rigging here, but rigging is important here).
4. Compatible UV maps (textures).
5. Compatible vertices (mostly this is what you get from morphs generated by bodyslide, but you also tend to see it in the first alternate bodies that show up after a game release).
6. Compatible skeletons
7. Compatible with game features (like dismemberment)

Game specific modding tools tend to use approaches that let you re-use existing parts of the game (like skeletons and rigging and textures).

Traditional tools for Artists tend to focus on composition and generality and aren't specific to any game.

So this tend to be tough to talk about. Someone says "use ____" and they just know that something went wrong for someone who didn't do that, and they aren't even thinking about what makes data structures compatible or incompatible with each other.

Meanwhile, eager users try things out and they don't work, and they don't have the vocabulary to talk about any of this and they don't understand what they need to be asking to sort through any of this.

We get a lot of grumpy people from all of this -- it's kind of frustrating.

I should probably do my part and define the above jargon well enough that people can understand it (and/or correct me if I have misunderstood something important -- which might be likely for some animation data structure related issues some shader issues, etc. etc.)

Still... when people talk about compatibility, sometimes you can figure out which kind(s) of compatibility they meant. Other times... not so much...

Worse, when reading guides: it's very rare that the guide author will go into details about which forms of compatibility guided them. Instead, it's often "this worked for me" and you are on your own for figuring out anything relevant to what you are trying to do.

-----------------------

Imagine this, though:

A tool sort of like BodySlide, but without the sliders, without the groups, basically without any of the new abstractions it introduced.

Instead, this tool would fit clothes to a base body. (You would still use BodySlide or similar mechanisms to build the base body.)

The way it would work would be:

The clothes were built for a reference body (not perfectly, but that's the starting point). The program would extract approximate normals from the base body (and maybe the normal map) and use that to build a normal (and a distance along that normal) for each clothing vertex.

Then it would wrap the clothing around the new body using that information.

This would run into a few problems (that would need some experimenting to see if these can be solved nicely):

• Rigging near joints -- what is the relevant normal for vertices in the armpit area. This might need heuristics or might even need manual work.
• Bogus vertices -- lots of bethesda clothes are body replacers and they have cheap blocked off regions on skirts. These are easy to detect (they cut through the body) but can we build a decent heuristic alternative? (Like a dark shadowed region for the inside texture up to the waist?)
• Rigging in general -- Clothes vertices often will not match body vertices for either the before and after body. And the clothes might be buggy -- is there a good heuristic for detecting and fixing rigging so that the cloths stay roughly the right distance from the skin on the new body? (Loose floppy clothes, and clothes which don't cover joints would probably be easiest here.)
• Clipping -- Bethesda clothes would often clip into the body if they weren't body replacer clothes. But that's easy to detect and they could be pushed along the vertices until they are outside the body.
• Curves -- all of these are really collections of triangles, but they eye wants the bodies to be curvy. I think you could use the catmull clark subdivision algorithm to generate a high poly surface to use for calculation purposes. An important issue here is that triangle vertices on the body shouldn't poke through faces on the resulting clothes. Skintight clothes need their triangle vertices to line up (along the normal vector) with vertices of the body. So for skin tight clothes (or clothing components) it's almost like you're really trying to wrap the texture over the body.
• Did I mention rigging?