Reminds me of this post by Drew DeVault. I don’t get why they are so focused on supporting people that tie things up rather than the ones building the foundations. Turning people that build the software that empowers your hardware away from you isn’t a sustainable strategy.
Developing software and hardware together makes a lot of sense - it’s just the way they were doing it (fund random distros based on what kind of device you bought)), and the way they’re doing it now (give Manjaro everything) seems backwards for actually doing the core work to make everything usable - no one’s doing that. You get the worst of both worlds (single OS target vs. many OSes with none treated as special) now.
I do think that giving some separation in software and hardware development makes sense. For one, this way avoids creating knowledge silos and incentivizes people to actually document stuff. Developing software and hardware as one, leads to monoculture, and I don’t like monocultures. The most devious thing Microsoft has ever done was convincing hardware vendors writing the drivers themselves, instead of properly documenting their hardware. Would be a shame if something of that kind would happen with Pine64 and Manjaro.
Partnering with an OS to provide a decent out of box experience is a good idea (though controversial amongst people who want it to be a blank slate) - but Manjaro’s lack of professionalism and lack of mobile focus (versus say, pmOS) seems questionable.
The boot loader problems the author describes are also real, and not just an issue for distro maintainers/porters. My PBP has been soft-bricked a handful of times — and is now completely dead barring destructive SMT-soldering surgery I’m not prepared to do — because of the absence of reasonable firmware and boot loader management tools.
I love that PINE64 has pushed the envelope on affordable, hackable, non-x86 hardware. I’ve bought many of their products as fun hacking devices, and expect I’ll continue to do that.
What I won’t be doing is using their hardware for any projects where, “I bricked it during a routine OS update” isn’t an acceptable obstacle. So: home media player, sure; small office firewall or backup server, no way. Anything more critical or hands-off than that, forget it.
An unfortunate symptom of Pine stuff (at least the stuff that runs Linux - the iron and PSU seem pretty OK from what I hear) being nothing more than tinker toys for engineers, and a self-reinforcing cycle of sloppiness that prevents it from achieving any more than that. I’ve heard so many stories about weird design quirks with power management or low quality NAND that really affect the experience.
I know my PinePhone is slightly infuriating - the default phosh loadout is annoying. On the software side, one example: You try swiping down from the top for notifications or settings? No, you have to tap - and the welcome app that appears on startup tells you about this and links to a two year old GitLab issue. (I want to try Plasma or god forbid, sxmo, but swapping DEs without cleaning it up with a reflash seems unfortunately annoying.)
Good motivation. I especially like how seriously you take your position as a representative of postmarketOS and that you think preserving the credibility of that project is more important than some short-term-isms. Thank you for making good decisions.
This looks like a microcosm of the wider F/OSS ecosystem. Back when I started engaging with F/OSS, my first taste was RedHat Linux 5.something, which had a whole bunch of X11 window managers and also included KDE BETA4 (which installed in /opt, if I remember correctly - somewhere non-standard to make it clear that it wasn’t officially supported). At the time, almost all of the software that ran was cross-platform and a lot of it was not developed primarily on Linux: it was built on BSDs, on Minix, or (very often) on proprietary UNIX.
Gradually, a load of things shifted from being ‘open source $THING’ or ‘*NIX $THING’ to being ‘Linux $THING’ and increasingly to ‘$THING that works on Ubuntu’.
When Linux was starting to gain inroads against Windows, software diversity was one of the big selling points: monocultures are bad because they’re vulnerable to widespread attacks (a Windows vulnerability could be exploited by a worm to take over all of your infrastructure) and because depending on OS-specific features means that you’re vulnerable to that OS evolving in a direction that you don’t like. Now that message is rare and a lot of folks advocate for Linux everywhere, and even within the Linux world, of running their application in a container with exactly the same base libraries, libc, and so on on Linux so that there’s no diversity at all.
I find this very sad. I think the F/OSS ecosystem is a great source of disruptive technologies precisely because the diversity makes it easy to try something wildly different from what everyone else is doing and still have a decent chance of being able to reuse existing things.
So what do you have to do to install another distro on a PineBook Pro? Do you still end up with Manjaro branding on the boot screen?
Regardless, if it’s true that you can’t just install a distro by popping in a flash drive “much like you can on any other laptop,” that is really astounding for a project which claims to care about open hardware.
Reminds me of this post by Drew DeVault. I don’t get why they are so focused on supporting people that tie things up rather than the ones building the foundations. Turning people that build the software that empowers your hardware away from you isn’t a sustainable strategy.
Developing software and hardware together makes a lot of sense - it’s just the way they were doing it (fund random distros based on what kind of device you bought)), and the way they’re doing it now (give Manjaro everything) seems backwards for actually doing the core work to make everything usable - no one’s doing that. You get the worst of both worlds (single OS target vs. many OSes with none treated as special) now.
I do think that giving some separation in software and hardware development makes sense. For one, this way avoids creating knowledge silos and incentivizes people to actually document stuff. Developing software and hardware as one, leads to monoculture, and I don’t like monocultures. The most devious thing Microsoft has ever done was convincing hardware vendors writing the drivers themselves, instead of properly documenting their hardware. Would be a shame if something of that kind would happen with Pine64 and Manjaro.
The project has posted a response.
Further response by Drew DeVault
Partnering with an OS to provide a decent out of box experience is a good idea (though controversial amongst people who want it to be a blank slate) - but Manjaro’s lack of professionalism and lack of mobile focus (versus say, pmOS) seems questionable.
The boot loader problems the author describes are also real, and not just an issue for distro maintainers/porters. My PBP has been soft-bricked a handful of times — and is now completely dead barring destructive SMT-soldering surgery I’m not prepared to do — because of the absence of reasonable firmware and boot loader management tools.
I love that PINE64 has pushed the envelope on affordable, hackable, non-x86 hardware. I’ve bought many of their products as fun hacking devices, and expect I’ll continue to do that.
What I won’t be doing is using their hardware for any projects where, “I bricked it during a routine OS update” isn’t an acceptable obstacle. So: home media player, sure; small office firewall or backup server, no way. Anything more critical or hands-off than that, forget it.
An unfortunate symptom of Pine stuff (at least the stuff that runs Linux - the iron and PSU seem pretty OK from what I hear) being nothing more than tinker toys for engineers, and a self-reinforcing cycle of sloppiness that prevents it from achieving any more than that. I’ve heard so many stories about weird design quirks with power management or low quality NAND that really affect the experience.
I know my PinePhone is slightly infuriating - the default phosh loadout is annoying. On the software side, one example: You try swiping down from the top for notifications or settings? No, you have to tap - and the welcome app that appears on startup tells you about this and links to a two year old GitLab issue. (I want to try Plasma or god forbid, sxmo, but swapping DEs without cleaning it up with a reflash seems unfortunately annoying.)
This is fixed on Phosh 0.20, which now supports gestures.
Good motivation. I especially like how seriously you take your position as a representative of postmarketOS and that you think preserving the credibility of that project is more important than some short-term-isms. Thank you for making good decisions.
This looks like a microcosm of the wider F/OSS ecosystem. Back when I started engaging with F/OSS, my first taste was RedHat Linux 5.something, which had a whole bunch of X11 window managers and also included KDE BETA4 (which installed in /opt, if I remember correctly - somewhere non-standard to make it clear that it wasn’t officially supported). At the time, almost all of the software that ran was cross-platform and a lot of it was not developed primarily on Linux: it was built on BSDs, on Minix, or (very often) on proprietary UNIX.
Gradually, a load of things shifted from being ‘open source $THING’ or ‘*NIX $THING’ to being ‘Linux $THING’ and increasingly to ‘$THING that works on Ubuntu’.
When Linux was starting to gain inroads against Windows, software diversity was one of the big selling points: monocultures are bad because they’re vulnerable to widespread attacks (a Windows vulnerability could be exploited by a worm to take over all of your infrastructure) and because depending on OS-specific features means that you’re vulnerable to that OS evolving in a direction that you don’t like. Now that message is rare and a lot of folks advocate for Linux everywhere, and even within the Linux world, of running their application in a container with exactly the same base libraries, libc, and so on on Linux so that there’s no diversity at all.
I find this very sad. I think the F/OSS ecosystem is a great source of disruptive technologies precisely because the diversity makes it easy to try something wildly different from what everyone else is doing and still have a decent chance of being able to reuse existing things.
So what do you have to do to install another distro on a PineBook Pro? Do you still end up with Manjaro branding on the boot screen?
Regardless, if it’s true that you can’t just install a distro by popping in a flash drive “much like you can on any other laptop,” that is really astounding for a project which claims to care about open hardware.