I’m not certain exactly why the Blue Oak Model License is not OSI- or FSF-approved, but here’s one mailing list post on OSI’s license-discuss list from one of its creators at about the time of its release: <https://lists.opensource.org/pipermail/license-discuss_lists...>. As part of this complaint about the terrible pain of the process (and it looks from other things like Luis had been involved with OSI as an invited expert for years, so he was familiar with this process from the inside): “I have no current plans to submit the license primarily because I am too busy to have a massively inefficient discussion on license-review. But I'd be willing to make that time commitment if I thought it was helping prototype something new and better.”

The people who made Blue Oak Model License have been involved with the open source community and with OSI. I think that the fact that it’s not currently OSI approved is more an indictment of OSI’s broken process, rather than anything about BlueOak-1.0.0.

CC0-1.0 is an example of a license that was submitted to OSI but withdrawn after a week because the mailing list ruined everything, people getting all het up about implications of the fact that it explicitly declared patents out of scope; <https://en.wikipedia.org/wiki/Creative_Commons_license#Zero_...> has a decent summary with links. Mind you, plenty of OSI-approved licenses make no mention of patents, and as a non-lawyer I find it difficult to imagine that any court would ever actually interpret a “patents are out of scope” any different from a non-mention of patents—the only way I can see it doing so is if ignorance actually is a defence. And of course they’re out of scope in a public domain dedication, a patent grant can only belong to a license, which a public domain dedication is not. (The fallback license could include a patent grant, but it would only apply in places where the public domain dedication failed, and would thus be futile. No, any patent grant needs to be a separate document.)

Well, when Creative Commons Zero lingers not OSI-approved, something’s at least a little broken about the process. When Unlicense can then be finally rubber-stamped in 2020 despite being acknowledgedly badly-drafted (quite apart from its atrocious name where “Unlicensed” means the exact opposite of “unlicensed”)—and Unlicense simply makes no mention of patents, and so should be in the same boat as CC0—something’s rotten about the process.

Yeah, it might be nice if BlueOak-1.0.0 were approved by OSI, but I personally am not going to fret much about it. As it stands, I’m currently tending to release my Rust libraries trial-licensed (two is dual, three is trial!), BlueOak-1.0.0 OR MIT OR Apache-2.0 because of the greater recognition of MIT and Apache-2.0 and their prevalent use in the Rust ecosystem. If I were making Node.js stuff, I’d go BlueOak-1.0.0 OR ISC because of ISC’s popularity in that ecosystem.

—⁂—

As for the GPL-compatibility question: their answer is very deliberately worded and perfectly accurate. They believe BlueOak-1.0.0 to be GPL-compatible (as I, a layman with his head screwed on fairly well and emphatically not a lawyer let alone your lawyer, am perfectly content to say it is also), but do not have the standing to declare it factually so without potentially incurring liability if a court ever decides it disagrees (for courts are the final arbiters of all licensing matters), and FSF have established themselves as the general arbiters of GPL-compatibility, so it would be somewhat impolite to declare something GPL-compatible without their approval of the matter.

—⁂—

I want to clarify that this comment is based only on my casual observations from outside any of these things, and is my opinions of something I don’t know all details of. I could easily be wrong on any point, and easily have erred in any conclusion from any point.