In my opinion Fabric is an exciting new modloader that may just be able to not get steamrolled by Forge (surpassing forge would be virtually impossible, but being able to thrive in the shadow of it is a great accomplishment).

The MC mappings are great, the use of Mixins makes APIs flexible, and the focus on utilizing interdependency seems interesting. The lack of Forge means a large (but modular this time) library would be very useful. For example in an unfinished Hooked port I implemented a system for storing data on entities, similar to capabilities in terms of storage, but without the focus on providing interfaces for items/energy/whatnot.

This doesn't need to happen soon or may not at all, but I think it's something to look into. The fabric branch was going great until I started bikeshedding about the merits of different vector APIs and stopped actually working on them.

The things I think would be worth porting over would be

• GUIs (I've had several people ask me about this on fabric already)
• Serialization with Prism (I'll need to get Prism working before we could add this one)
• Math (Potentially rewriting the currently janky system, having separate Vec2i and Vec2d classes and stuff)

It probably wouldn't end up being called LibLib though, because having two different versions with different features for different modloaders would become confusing real quick. In an ideal world we would be able to abstract LibLib away from forge or fabric so we can support both. Emphasis on ideal.

With the support of other interested parties I feel like this might be doable.

Further discussion will be here: https://discord.gg/xPQAKf5