NASA and open-source software - public domain

You're right, if there's no copyright AND there's no other constraint (e.g., patents & trademarks), you can insert the no-copyright work into a work with a copyright. So it's not a "problem" in the typical sense, because you CAN legally do it.

The usual objection involves government folks being worried that they'll be re-paying for software the government already paid to develop. OSS licenses - even permissive ones - typically *do* impose some requirements, typically an attribution requirement. If the software has to reveal that it's written from XYZ, a user can find XYZ. You can't impose *any* requirements on a work without copyright, including an attribution requirement. In my opinion, the real solution is that government people should do market research before they buy anything, as already required by regulation. Market research in most cases would quickly find them alternatives. But government employees are typically overworked, and market research is often skimped as a result. That still doesn't really excuse it though; don't pay a lot of money before looking at your alternatives.

The blog post "How to apply an Open Source License to a US Government Work" by Dan Risacher https://risacher.org/jfdi/2014/07/how-to-apply-an-open-so... does an awesome job discussing this. Basically:
1. Just release as public domain.
2. Create a joint work that's partly public domain.
3. Rely on the fact that "no copyright" only applies within the US; copyright still applies outside the US.

There are some weirdnesses in this case. Some organizations (e.g., Apache Software Foundation) require organizations to sign statements involving copyright, and government lawyers get all confused. Historically organizations would make exceptions for the US government, but many foundations aren't familiar with the peculiarities of the US government. There's still a copyright outside the US, so I think they can simply sign with a clean conscience, but I'm not a government lawyer so they don't need to listen to me.

The OSI's position is that "public domain is not a license". That is true in a sense, but misleading. There's a license, it's just that it's granted by the laws of the relevant nation instead of being a separate document. A bigger concern by the OSI - and one that *is* reasonable - is that simply not having a copyright doesn't mean that it's legal to use as OSS, because it could still be covered by patents and/or trademarks. Trademarks are easily dealt with, but patents aren't. If software is covered by patents, then the lack of copyright is irrelevant, and in this case the patent permissions granted by typical OSS licenses don't work (like the implicit ones in MIT and GPL 2.0, or the explicit ones in Apache-2.0 and GPL-3.0).

My personal opinion is that this is mostly a NASA "problem", in the sense that NASA's lawyers think it's a problem even though most of the rest of the US government doesn't think it's a problem to release code as public domain. Many programs have been released without copyright (e.g., expect), and the sky did not fall.