A shadow lies upon all BSD distributions

Or so it seems. I’ve written yesterday about the troubles that forced me to get the Gentoo/FreeBSD stages out of the mirrors, to feel safer and to avoid issues to the Gentoo Foundation that would get bit if there was a problem; today I was invited to join #gnu by mattl on Freenode (who was in turn invited to join #gentoo-dev by christel), and there I talked with ams (Alfred M. Szmidt), who agreed with me that the clause is way too vague (what would be considered “advertising material”? a poster? a booth at an expo? a website? a document explaining the installation procedure? an article on a magazine?), and to be safe we’d have either to get the permission from all the entities involved there, or list all the acknowledgements for all the entities (at least 110 if I have to depend on the quick’n’dirty grep I posted yesterday, but probably a lot more after seeing the files directly, see later on in this post). Both strategies are difficult to apply on both short and long terms.

But this is not just a problem for us, as the title of this post already made you suspect. All of the *BSD-derived projects took some way or another code licensed 4-BSD that is not under copyright of UCB, that would then require them to provide the acknowledgements on all the “advertising material”, whatever that is. And I’m pretty sure most of the *BSD projects have something that can be considered advertisement even to the stricter of the meaning.

So even if the situation is nasty, and not easy to cope with, and not even quick to deal with, we have one advantage: the same situation is true for other projects, and thus it’s well possible that we’ll be able to find all the 4-BSD licensers and get them to change to 3-BSD, or replace the code with cleanroom implementations that would be licensed under a saner license.

Javier is taking care to contacting FreeBSD Foundation, while Timothy took care of NetBSD, I’ll probably try to find a way to contact at least some of the copyright holders, but to do that I have to be able to reduce the copyright headers into a human readable form without all the boilerplate noise around it. To do that, I’ve written a short stateful parser in Ruby that produced me already an interesting log: a lot of files that resulted from the grep assigned to NetBSD project, are actually assigned to various companies or developers that then assigned it to NetBSD project, which makes the situation even more troublesome.

I suppose now people can actually say that my work destroyed the very foundation FreeBSD is based off :/

Edit: It seems like *BSD projects are not the only ones that are involved in this debacle, as also Open Darwin includes at least some header files that are licensed under 4-BSD, see for instance libc_private.h; I’ll have to check if the include directory of Mac OS X also contains 4-BSD licensed code, because I don’t remember seeing the acknowledgements on any Apple advertisement.

Why did I have to discover this hornets’ nest myself? :§