Right, so you might have, for example, FreeBSD, which as distributed by the FreeBSD developers is free software, but which can also potentially be used in a proprietary downstream product, which is not free software. The FreeBSD developers did not deny users any of the four freedoms. They just allowed other people to potentially do so.

This Venn diagram is FSF's view of the taxonomy:

https://www.gnu.org/philosophy/categories.en.html

I was wrong indeed, thanks for the precision :-). So:

Permissive licenses are free in the sense that they allow making free software (but they also allow making non-free software).

GPLv3 enforces the complete derived work to be free. LGPLv3 says that the particular library should stay free.

GPLv2, LGPLv2 and MPLv2 say that the source code (either of the whole derived work or just the library) should stay free, but tivoization allows making the product non-free.

Is that about right?

That matches my understanding. Some people might potentially only say "tivoized" (or a more explicit "incorporated in a product that prevents the user from modifying it in practice" or "incorporated in a locked-down product" or something) rather than "non-free" for a hardware product, but I guess "non-free" could also make sense for a hardware product.

That's spot on.

That was what I meant initially: in HN there's a strong opinion against copyleft, not necessarily free software.

I get why companies can make more profit in a world where most software is under a permissive license. I just don't really see how it benefits users and, more importantly, developers as individuals.

To me, writing software in my free time and open sourcing it under a permissive license is shooting myself in the foot. I did the work for free, it may as well benefit me as a user. At minimum it should be MPLv2.

Please note that "Tivoisation" isn't what Tivo actually did and both the GPLv2 and GPLv3 ban "Tivoisation" but allow what Tivo actually did. "Tivoisation" as it is popularly known refers to blocking the running of modified GPL code, while what Tivo actually did was block running their proprietary software (their UI etc) on top of modified GPL code (here Linux). Both GPL versions block the former while allowing the latter. At least according to Software Freedom Conservancy.

https://sfconservancy.org/blog/2021/mar/25/install-gplv2/ https://sfconservancy.org/blog/2021/jul/23/tivoization-and-t... https://events19.linuxfoundation.org/wp-content/uploads/2017...

That's interesting, I remember thinking that the original notion of Tivoization did apply to what Tivo did (including what you describe).

I wonder if this is a point of disagreement between FSF and SFC -- I know they now have several disagreements.

In a similar time frame, I criticized attestation features in trusted computing because they would allow (in fact one of their main purposes was to allow) network services to allow only certain software configurations to interact with them. And I thought I was talking about a similar concept!

Thanks for sharing those links.

You may be thinking of SFLC, their disagreements with FSF are much larger. SFC's disagreements with FSF are mostly related to keeping RMS around.

Attestation is indeed a big problem today, especially around Android devices and proprietary apps that won't run on libre Android distributions. The attestation feature of WebAuthn could also become a problem, but that is somewhat mitigated by the Apple passkeys not being attestable.