Excellent interview with Hyperbola co-founder on why they are making a libre BSD kernel, and other stuff
- Vous devez vous identifier ou créer un compte pour écrire des commentaires
https://itsfoss.com/hyperbola-linux-bsd/
Andre, the co-founder of Hyperbola, makes a number of really good points in this interview. All of his arguments are from the perspective of making a distro that is more compliant with GPL3:
1. The Linux kernel is rapidly proceeding down an unstable path:
(a) "...it’s including the adaption of DRM features such as HDCP (High-bandwidth Digital Content Protection)
(b) "the Linux kernel is no longer getting proper hardening"
(c) "the interest in allowing Rust modules into the kernel are a problem for us, due to Rust trademark restrictions which prevent us from applying patches in our distribution without express permission. We patch to remove non-free software, unlicensed files, and enhancements to user-privacy anywhere it is applicable"
2. Why he chose to fork the openBSD kernel:
(a) "new system calls, including pledge and unveil which adds additional hardening to userspace and the removal of the systrace system policy-enforcement tool"
(b) the openBSD kernel is also "known for Xenocara and LibreSSL, both of which we had already been using after porting them to GNU/Linux-libre"
(c) "LibertyBSD has been working on liberating the OpenBSD kernel, which allowed us to use their patches"
3. What percentage of the OpenBSD kernel code falls into the non-GPL zone? How will they make it a libre kernel?
(a) "around 20% in the OpenBSD kernel and userspace [are currently non-free]"
(b) "The non-free firmware blobs in OpenBSD include various hardware firmwares. ... These blobs may contain vulnerabilities or backdoors in addition to violating your freedom, but no one would know since the source code is not available for them. They must be removed to respect user freedom."
4. They have forked the Palemoon "UXP" browser code to create a suite of web applications.
5. They are forking pacman and openRC to work on BSD
6. Planning to port the following file systems: "BTRFS, JFS2, NetBSD’s CHFS, DragonFlyBSD’s HAMMER/HAMMER2 and the Linux kernel’s JFFS2, all of which have licenses compatible with GPLv3"
Lots more good stuff. Quite an interesting person with a very strong message!
the interest in allowing Rust modules into the kernel are a problem for us, due to Rust trademark restrictions which prevent us from applying patches in our distribution without express permission.
That makes no sense to me. Trademarks policies only apply when one uses something that identifies a product (its name, its logo, etc.). As far as I know, the Linux project does not plan to change the name of the kernel for RustyLinux (or something similar) and the code written in Rust would necessarily be under the GNU GPL, granting the same freedoms (including modification and redistribution) to its users as the rest of the kernel.
Even if Linux was renamed RustyLiux (maybe after asking permission to the Rust team) and would adopt the same trademark policy as Rust, it would still be under the GNU GPL and any developer could distribute a patched version under another name. Exactly as Trisquel does with Firefox that becomes Abrowser.
"around 20% in the OpenBSD kernel and userspace [are currently non-free]"
It is not what Andre said. That would be wrong. He said "around 20% in the OpenBSD kernel and userspace [are not GPL compatible or non-free]" and complements: "Mostly, the non-GPL compatible licensed parts are under the Original BSD license, sometimes called the “4-clause BSD license”". It is indeed incompatible with the GPL. But it is a free software license anyway.
Rewriting much free software so that it becomes GPL-compatible looks like a a lot of work (done without the help of BSD developers, since the stated goal is "future development under GPLv3 and LGPLv3", i.e., developments that the BSD projects would not profit from). That effort could be put elsewhere. Anyway, the HyperbolaBSD project is a free software project: I wish it success.
Rust is on here: https://www.mozilla.org/en-US/foundation/trademarks/list/
It inherits the same problem that Firefox and all of the other things on that list do, namely that Mozilla's trademark policy tries to go beyond the usual trademark realm where it "identifies a product" to include distribution restrictions. Distributing is supposed to be the realm of copyright, but Mozilla's trying to go there via trademark instead. Specifically that exact copies must be distributed at no cost (primarily conflicting with freedom #2 in the Free Software Definition for the making of exact copies.)
It would be helpful to have a re-branded version of Rust where exact copies can be distributed commercially.
"Even if Linux was renamed RustyLiux (maybe after asking permission to the Rust team) and would adopt the same trademark policy as Rust, it would still be under the GNU GPL and any developer could distribute a patched version under another name. Exactly as Trisquel does with Firefox that becomes Abrowser."
Distributing a modified version is using freedom #3. There's no dispute that freedom #3 still exists so that's not the point. The matter is over freedom #2; the making of *exact* copies (literally invoking cp or scp or rsync or whatever to give someone an exact byte-identical copy.) Not being permitted to do that commercially conflicts with freedom #2. To quote: "Freedom to distribute (freedoms 2 and 3) means you are free to redistribute copies, either with or without modifications, either gratis or charging a fee for distribution, to anyone anywhere. Being free to do these things means (among other things) that you do not have to ask or pay for permission to do so."
And currently you're not allowed to redistribute copies without modification when charging a fee (at least, without getting permission); and so only part of that criteria are met.
Fortunately being based on trademark allows for easy escaping: Someone can make a version that allows all of those criteria and that modified version can become the new upstream for the free software community to use instead of the original upstream. But the matter I was talking about wasn't the ease of escaping; but evaluating how the original upstream versions of Rust and Firefox and etc. themselves meet the criteria in the Free Software Definition which as you can see they fall short.
Distributing a modified version is using freedom #3. There's no dispute that freedom #3 still exists so that's not the point.
It is Andre's point. He is not talking about exact copies of Rust utilities (the compiler and so on) but about "applying patches in our distribution without express permission" to software that is written in Rust (not by the Rust project). So, you are here confirming what I was writing: that makes no sense; Rust's trademark policy does not impose the restriction Andre pretends it imposes on modifications. Not even on modification of the software written by the Rust project.
But there are freedom problems in Rust and the Linux kernel having code written in Rust then leads others to make use of it. But this problem is solvable by having a rebranded version of Rust.
I saw this a week or so ago, I look forward to this project even having an alpha, let alone a beta or stable even!
I will try it out on a computer i don't use often probably till its more stable.
(Like beta style stability to start with)
Anywho, it is a shame no one has had the guts to try this sooner. ;)
But maybe its for the better this way, a fully free bsd will be very tricky due to all the non-free nonsense within it... and we need someone who is actually competent to audit the whole damn thing to be able to make it fully libre. :)
> But maybe its for the better this way, a fully free bsd will be very tricky due to all the non-free nonsense within it...
Looks like the LibertyBSD scripts are still being maintained, so you could probably try LibertyBSD while waiting for HyperbolaBSD:
https://notabug.org/LibertyBSD
Small problem though,
https://notabug.org/LibertyBSD/libertybsd-scripts/issues/41
Apparently even libertybsd has a lot of non-free stuff in it.
Because even the userspace... heck the kernel itself also has a lot of non-free stuff in it. so there's that to be concerned with.
She actually gave up on maintaining it last time i checked.
But yeah libertybsd is more free then openbsd... one other problem though is its outdated. meh... I wait.
Yes but if I read the comments below, which you were involved in, a lot of the licensing of those files isn't actually a problem. Also, the scripts appear to be maintained up to current time on the git page I linked to. Do you have additional info about the scripts on that git page that I am not aware of? Did someone else take over the project once the original developer left?
Hard to say, I personally talked with the woman who was developing libertybsd. She told me she had stopped developing it following that, I tried to get her interested in working on HyperbolaBSD though. ;)
And as for any other info, that I do not have at this time.
From the interview with Andre and the page that you linked to, it looks to me like maybe the Hyperbola devs have taken over those scripts. But I don't know them and you do, so you would know better about what they are up to.
- Vous devez vous identifier ou créer un compte pour écrire des commentaires