The only entity that can change of ZFS license is Oracle and they obviously wouldnt do that.

They could rewrite all the code, and then change the license. Patents might still apply (but patents are short enough that I expect if any existed they have expired). However ZFS is a lot of code that is often tricky to get right. It will be really hard to rewrite in a way that the courts don't (reasonably/correctly) say wasn't a rewrite it was just moving some lines so you can claim ownership, but it is possible. By the time anyone knows enough about zfs that they could attempt this they are also too tainted by the existing code.

So of course they won't, but it isn't impossible.

I mean, bcachefs is basically the equivalent of rewriting all that code, without explicitly trying to be a clone. Same for btrfs

And how hard it is proves that zfs didn't make a bad choice in not trying the same. (though it would be interesting if either had a goal of a clone - that is same on disk data structures. Interesting but probably a bad decision as I have no doubt there is something about zfs that they regret today - just because the project is more than 10 years old)

It's supposedly the opinion of Oracle that the CDDL is GPL-compatible and that's the reason they won't do that.

I would not rely on the non-binding opinion of a company known for deploying its lawyers in aid of revenue generation

For context: https://sfconservancy.org/videos/2023-07-14_FOSSY-2023_RHEL-... (FF to 37:34)

That wasn't exactly the answer.

Yeah, I agree based in rewatching that I've either misrecalled the original material, or I got it from another source.

I agree that based on that source, it's more like "meh, we don't really care" (until they do)

Oracle didn't follow that with DTrace. They changed the license away from CDDL when they integrated it into Oracle Linux.