February 15, 2024 Categories: Interview

An Update on the HyperbolaBSD Project

Back in 2019, the dev team behind the Hyperbola GNU/Linux announced that they would be moving away from the Linux kernel. They decided to fork the OpenBSD kernel. Not only that, they planned to rewrite and replace all code that is not GPL v3 compliant. Back in January of 2020, I interviewed the HyperbolaBSD team for ItsFoss.com. I decided to reach out to them again to see how the project has progressed. I also included some questions from fellow nerds. Enjoy the interview.

Q: Back in December 2019, you announced that you would be moving from the Linux kernel to your own fork of the OpenBSD kernel (the Hyper Berkeley Kernel). It’s been over 4 years. How have things been progressing? When do you think you’ll have an alpha build available? When do you think it’ll be ready for day-to-day use?

A: Since December 2019, there were many things happened along the time. One of them was the COVID-19 pandemic, which not only affected our planning to begin the project and our lifes, but the whole world, with loosing of many lifes. For this reason HyperbolaBSD development begun exactly at 20th March 2022.

But in comparison to before, we have a better organization/planning how to develop it.

For the first version, our planning have 4 stages of development:

Pre-alpha:

Port the x86 system from binutils 2.17 to 2.34 (with HyperbolaBSD patches)
Port the x86 system from GCC 4.2.1 to 8.4.0 (with HyperbolaBSD patches)
Port the x86 system from GNU C99 to GNU C17 standard
Adapt code to be built under FreeBSD bmake

Alpha:

Write compatible code under Simplified BSD License to replace the non-free files in kernel and libc
Modularise kernel and userspace under the Unix modular design
Replace non-free tools/applications with fully libre third-party ones in userspace

Beta:

Develop hyperman (hard fork of pacman adapted for HyperbolaBSD)
Develop hypertools (hard fork of libretools adapted for HyperbolaBSD)
Migrate our build-server (Laminar) towards HyperbolaBSD
Port runit for HyperbolaBSD in order to replace the OpenBSD init
Develop runit init scripts to run HyperbolaBSD
Package the entire modularised HyperbolaBSD system with hyperman
Develop the HyperbolaBSD live image
Port Xenocara for HyperbolaBSD

Release Candidate/Final version:

Port and package software included in our repository named “extra” from Hyperbola GNU/Linux-libre to HyperbolaBSD, including also an audit of being really needed or just be seen as nice addition

Currently, we are very near to launch our pre-alpha version which will be ready for testing purposes only. There isn’t an exact prevision of the date for all of those versions, but we will let you know in our website’s news.

In other hand, the original HyperbolaBSD idea planned at December 2019 continues in the same way than before, however under a progressive migration by replacing all non GNU GPL-compatible code along the next HyperbolaBSD versions in the future. The difference here is about it will be replaced with new compatible code under Simplified BSD License, not the GNU GPLv3. We have decided for it in order to contribute for other BSD descendant operating-system projects. But for HyperbolaBSD side, also it will be useful as well, because we could incorporate GNU GPL code from other projects such as ReactOS, as well new code from scratch made by us in HyperbolaBSD.

By the way, I would to like to emphatize here that our plan never was just to move the Linux kernel to an hard fork of the OpenBSD kernel, but the develop an independent BSD descendant operating-system, based on hard-forked code from different BSD descendant operating-systems (mostly from OpenBSD 7.0), as a starting point for the development of the first HyperbolaBSD version.

Q: What unexpected problems have you encountered?

A: Currently we didn’t find critical problems, but yes how to port the whole code at the beginning stage of the development, with a different version of the C standard, as well binutils/GCC, because this part of the development requires do a drastic change.

Q: Have some problems turned out to be easier than you thought?

A: For now we didn’t find that possibility, but it could be probable along the development in this first version.

Q: In the HyperbolaBSD roadmap, you have an entry to remove non-x86 from the kernel and libc. It appears that initially you will only support x86 systems. Do you plan to add more architectures in the future?

A: We have plans to do it in a long-term, but there are other priorities to do first.

Q: In my previous interview, you mentioned that you “need C programmers and users who are interested in improving security and privacy in software”. I know from your site that you have disagreements with Java and Rust. Would you accept code from other languages?

A: A base for an operating-system should be most focussed on one language. Creating HyperbolaBSD can include C and / or C++. Regarding Java and Rust especially it is not possible to include code using them especially from the perspective of their trademarks, mainly in the redistribution restrictions. Trademarks are from our perspective one major trap when done very strict and not allowing further modifications even block exactly the freedoms libre software should grant for users and developers. We need to make a difference between abusive trademarks and others allowing freedom.

But this is a point we want to explain a bit more deep towards the example of Rust: They demand in the regulations within their “Logo Policy and Media Guide” only some minor possible modifications being allowed. Otherwise a complete rebranding of the whole programming-language is needed, resulting in fact towards a complete hard fork. A rebranding of a complete programming-language is different in comparison to a rebrading of an application for endusers, web-browser as example to be named. A rebranding cannot be done with a simple command and needs a complete review of all components, not only binaries like the compiler named different.

Furthermore Java following comparable strict, problematic trademarks and with its own history of security-flaws over the years is nothing we want to bother also. It would also distract further the focus if we would accept more code from other languages as it would also need porting, packaging and intense testing: Another point of distraction. So we stay strict and straight oriented on our roadmap and plans.

Question from the community:

Q: I realize this question comes off as condescending, but I mean it with sincerity. Why not take your considerable developer time and expertise and contribute to a project that already meets your stated reasons for leaving the Linux ecosystem like OpenBSD? Wouldn’t assisting that project contribute more to the BSD ecosystem then yet another project?

A: OpenBSD is already including software with freedom-issues and the ports are doing same. The decision for another BSD-descendant operating-system is also a decision about software-freedom. Sure Hyperbola Project follows a right different perspective as others and that marks the important point: Every new project adds another variety for a colourful software-landscape regarding libre software. The major aspect for Hyperbola Project, including HyperbolaBSD and also Hyperbola GNU/Linux-libre, is our focus on technical emancipation: We want to motivate users to rethink about false balancing to accept non-free interfaces and inclusion of questionable implementations. Giving full control over every aspect of the system users can be sure all parts do not include any unknown firmware-blobs and also audits for issues and flaws are way more easy.

Q: Do they belive in the idea of perpetual development of software - or do they perhaps instead hold that there is such a thing as “optimal software” - software that has achieved a sort of perfection or optimum and doesn’t need to be developed anymore? Does this belief of theirs, whatever it is, factor into their decision to commit to a ten-year development process?

A: That depends on the software itself and its operation area: Building an operating-system is different towards a generic application for the commandline or with a graphical user-interface. Software with intense security-focus will always need perpetual development, this includes also the kernel and the userspace. Other applications outside may reach a point being defined as “working perfect” and so frozen in development. So as long as we have support and can work with passion on HyperbolaBSD we will do that as Hyperbola Project will be always community-focussed. Anybody can use our development for a base to build as we have also neither the interest nor the need to follow too much implications from outside and can pronounce our own perspectives.

Q: What areas of the project do you think need some TLC that other areas get?? How can one contribute to the project?? Who to contact at the project?? Do you need beta testers?? If so how to sign up??

A: That depends on the individual perspective: Working on security-aspects, replacing questionable licensed and non-free code, removing non-free blobs and system-calls, there are many different task life cycles in need for attention. Also the on-going packaging and providing a working infrastructure for users doing ports. But exactly coming the last mentioned point: Contributions are possible on many ways like writing documentation, supporting the development for our essential packaging-tools towards our system or also granting support for users with porting and packaging. Therefore you can always get in touch with us through the forums as every active team-member can help. So yes, we also need beta-testers when entering the concrete phase. Reporting for this will be done through announcements in our forums.

Q: What HyperbolaBSD recommends for web development?

A: That depends foremost not on the operating-system itself and is more a generic question towards security: Web-development is a risky field. Neither NodeJS nor others like PHP or Java can be seen to be approved working in the perspective of trademarks and security itself. A recommendation is therefore near impossible while best would be to minimize the attack-surface. Complex web-development should be avoided generating a false perspective of being possible without flaws.

Q: What HyperbolaBSD recommends for desktop app development?

A: Same as before: The more simple the better especially in usage of frameworks. Qt has proven to be not fully working as for example including non-free Chromium partwise and licensing-model being not fully oriented towards libre software. GTK+ is going more into the direction being no longer oriented backwards-compatible and including more interfaces and frameworks making maintenance even more complex over time. While we can’t give a generic recommendation the perspective is the most important: More simplistic designed applications for desktop-environment with a clear set of tasks is way better than applications getting too complex. Besides we should always remember that “app” is misusage of the wording “application”, which is already the concrete statement about a straight focus Hyperbola Project is following.

If you found this article interesting, please share it on social media. You can also send a tip via Paypal.me or Liberapay. If you would like to comment, please use the contact form listed at the top of the page.