Not a good choice for a name, at first I though it was just another linux phone that would be useless for 90% of people.
Very cool project instead, hope this can lead the fondation for a 100% open source mobile OS.
Not a good choice for a name, at first I though it was just another linux phone that would be useless for 90% of people.
Very cool project instead, hope this can lead the fondation for a 100% open source mobile OS.
I think this intiative is spot on. I would describe current approach of 2 major OS vendors, Google and Microsoft as such:
Microsoft demands standardization at firmware level via UEFI, ACPI etc. because they bring OS kernel and userspace.
Google demands Linux API version and brings just userspace.
In theory Google approach better facilitates open ecosystem but each OEM treats Linux kernel as just a firmware blob so the end situation is actually worse.
On the PC we have standardized firmware while Android chases Linux API levels each release and thus undermines the whole ecosystem.
So they could do it for pixels and this open source firmware could be used by Graphene OS, for example?
Yes, though the future of GrapheneOS on Pixels after 10 is currently in question
The issue is that for the FSF, what they call "software freedom" is their number one goal. So what's likely to happen is that they create some kind of "deblobbed" firmware that breaks many features and security of the device, which Graphene OS will refuse to use.
I hope this project will be useful but am worried that they'll just make a shittier version of someone else's work like they did with e.g. Libreboot.
I don't know as much about phones as I do with computers but would I still be able to contribute in any way?
I'm trying to fund these projects. It might not be much, but but a little bit now and then might make a difference.
A dedicated community has been created for this, see https://lemmy.ml/c/librephone (I just saw this community and joined myself)
Marketing nerds hate to see a FOSS project name roll up 😎😎
Don't like LibreThing? Just use OpenThing. Or FreeThing. Or PeerThing if you're weird.
I will jump immediately no second thoughts once any open phone comes with a good stylus.
`+ e-ink variant o.o if possible?
If someone comes up with full colour reflected e-ink with refresh rate better than 24 fps, absolutely.
Theres the Daylight Computer, that has a 60fps display which is "like e-paper". (eink is the propietary name, epaper is the general name of the tech)
I wish this tech could receive more funding for research and development.
Genuinely curious, what on earth do you need a stylus for? I've never really understood the use case
Taking impromptu notes is what I use it exclusively for. Also to draw quick diagrams and such. The diagrams drawing helps especially.
Just word note taking can be done with any old notes app. But sketches and diagrams. That's where I use it most.
Oh fair. Diagrams makes a lot of sense
Yup. Try drawing a diagram and arrows with lines beginning and ending exactly where you want it intuitively with your finger. Not too mention you don't need to unlock your phone, navigate to an app to do it. Just pop out your pen and go. No need to keep anyone waiting if you're in the middle of discussing something.
Also one of the reasons I was looking to get a fold because the screen real estate was lacking. But they took away pen support in 7 so what the heck.
I hope it goes well.. but imo firmware blobs in LineageOS shouldn't be the highest priority.
What should be the highest priority in terms of being able to use a fully FOSS phone? I feel we already have FOSS alternatives for most things that make the phone usable.
Perhaps free hardware might be top, but the S in FSF is Software and being able to avoid non-free software would be close 2nd, imho.
from their article it seems like they're planning to reverse engineer firmware blobs for a phone that already runs a mostly free android distro (LineageOS). I know the importance of free firmware and stuff, but honestly, that phone is freer than most phones in the market. My opinion is that they should invest(?) in non-Android linux mobile OSes instead.
I think it makes sense to choose a phone that's freer than most phones as a start, then fill up the gaps to make it fully free. As things are all phones have blobs (even non-Android ones).
I expect non-Android linux mobile OSes can also benefit from the reverse engineered firmware.
I get what you mean, but may be this is a start. I guess, this work can help them understand phone architecture better and the community can build upon their work to improve non-Android Linux mobile distros. However, I think this will help projects like replicant, droidian and grapheneos.
How open source can they go? Is there a capable open source processor without propriatary backdoors yet? Will it use RISCV?
Is this just an effort to replace binary blobs, or to actually support usable Linux phones or custom roms...? Are those kinda the same thing in this case?
If it's the latter this might be the first thing the FSF is actually doing during my time with foss that I'm excited about.
If it's just that they hate binary blobs, that's great and all but I do kinda wish they'd focus on stuff that has a way bigger user impact 😅 I think they got really hung up on binary blobs with libreboot or whatever it was also
I've worked with Qualcomm SoCs a long time ago and, from my experience, the binary blobs ARE the biggest hurdle to the true Foss phone. Google is the most to blame, IMO, but also the rest of traditional OEMS of SoCs. They basically found the way to circumvent the OSS nature of Linux, which is why even though android is based on linux, the actual product and ecosystem looks nothing like regular Linux. What Google allowed with Android architecture, particularly with their HAL subsystem is force a layer in between native Linux device interface and Android system, so OEMs use that to implement whatever proprietary peripheral (device, sensor, etc) purely in userspace, rather that just as a kernel module. The kernel module is then just a userspace/kernelspace adapter, and everything is handled in the user space. This then means you dont have to have an open source driver, as it is not a part of kernel, and you just lock your driver into a binary blob. And in case of Qualcomm and I assume other oems, everything is just a binary blob. All sensors, microphone, GPS, modem, EVERYTHING. Yes, you can boot a basic Linux kernel, but no other functionality will work. If you had access to the blob source, you'd be able to fix, update and use a newer kernel versions eventually. HAL is technically not the cause of the problem, but it's certainly an inspiration and almost a blueprint for how to lock down your hardware.
I expect it's a combination of all the above in some sense. They state they want to build on LineageOS (an Android variant) and replace its binary blobs, I expect the result would be a new custom ROM targeting specific compatible hardware with the goal of ultimately supporting usable phones working on fully Free Software.
What it's not is the creation of a libre hardware phone. I don't think they are working on hardware, at least not anytime soon. Also if by "Linux phones" you mean non-Android based, that's not necessarily a requirement (given that they mention LineageOS), but I expect regardless the kernel will be Linux without the blobs and it's entirely possible that they add support for installing their firmware on those "Linux phones".
I do kinda wish they’d focus on stuff that has a way bigger user impact 😅
The thing is that technically we already have fully usable FOSS software at that user level. Using for example LineageOS with F-droid as the only app store already gets you there. Whereas, ensuring your phone is not spying you or having some malicious functionality on the hardware/driver level is something that currently is simply not possible.
The FSF has always been doing the thankless job of championing for the things that are harder and less rewarding to do, but that will advance software freedom most for those who do seek it. Even when that thing is not necessarily the most popular/mainstream. I feel this has more of an impact in software freedom than, say, if they were to reinvent the wheel just to have their brand attached to it, and/or provide a slightly different UI to do the same thing other FOSS software already does.
Gotcha, thank you very much for filling me in, I really appreciate it!
I do think there's value in ensuring everything low level is open source, but at least with linux phones none of the higher level stuff is really in a good state yet 😅
Its nice to hear this will be an improvement to custom ROMs. Though the future of custom ROMs is feeling a little shaky with the stuff Google has announcing lately 🥲
Awesome!
Having said that, please start with an existing open source project so you won't have to start from scratch and still be nowhere 10 years from now
That’s amazing! Phone are, IMO, the toughest frontier of privacy and autonomy. Industries have us well locked in their ecosystems
Let's see where this leads. They have not managed so far to get GNU/Hurd production ready in the 35 years it's been in development.
HURD is not in development by FSF. It's a side project by professor Samuel Thibault in his spare time.
It sounds like they are basically doing the same thing Replicant does (they even mentioned Replicant), but based on LineageOS right now, and with enough resources to hire someone to work on reverse engineering the proprietary bits.
Not the most glamorous endeavor, but a cool and necessary project. Pretty pragmatic too, focusing on Android instead of mobile Linux.
They should support GrapheneOS rather than try and make something "more free" from LineageOS.
Whaaat? Drop the OG free Android distribution that has already supported hundreds of phone models across decades for a Google locked hardware with unknown future support from a vendor just announcing locking down everything they can? What is your logic?
Logic is simple: Actually have something 10 years from now.
If google restricts access to its os, like they have already started, all you'll have is pixel up to 10/11 still supported 10 years from now. They've already started by no longer providing device trees in aosp for their phones, so graphene has to work harder to obtain them now. Whereas if you work on lineage, you potentially have a greater number of vendors and potentially new ones ready to open up to draw in new userbase.
Absolutely that's an issue. But we're not talking about that here. We are talking about a base os from where we can progress. I don't care much for Google phones, even though I must admit they are nice phones. Google can not un-opensource AOSP. They can, to a certain extend, stop open sourcing changes, but that's about it. Doesn't mean we need to follow their os. Also doesn't mean we can't, slowly but surely, develop Android to be more of a Sandbox ontop of a newer Linux kernel than it is right now. Utopian, I know, but if Google stops AOSP developmentz what would we rather do?
No, I mean, both graphene and lineage are based on aosp. But graphene supports only one vendor. Lineageos supports many, including google. Why invest in a vendor-locked os and risk loosing it all? I think lineage is a lot more logical choice. And I'm currently running Graphene on a pixel 8, after pixel 7.
In case anyone is misunderstanding, they explicitly say this is not a new phone or piece of hardware at all, it is simply a project (and for now, more of an investigation than a project creating actual deliverables) into the scale and scope of closed source binary blobs being used on phones, so they can start work to address them.
It's an important and necessary project, and I support the FSF in most of the things they do, but if you're picturing them riding heroically to the rescue by Christmas with a new phone-of-freedom they're going to sell to you, it's a very very VERY long way from that.
But on the positive side, someone is finally attacking the last obstacle to using our expensive hardware however we want to: device drivers. I'm hopeful.
Spot on. There are substantial incentive of each company wanting to keep user on platform they have control over, if not massive for revenues from appstore tax, advertising, etc.
I think it suffices to say that it is becoming a primary goal selling you their phone (I don't have data or evidence for this, just a feeling) -- to get you to use their system and stuck with their advertisement slot or whatever every time you unlock your phone.
Damn. They're basing it all on LineageOS. I was hoping they'd make it a GNU/Linux phone.
I hope this isn't overreach given the amount of secret crap in phone hardware. Plus phones connected to mobile networks are inherently invasive. RMS refuses to carry one for that reason.
Very ambitious, hope something comes of it. The focus on enhancing (?) compatibility between Lineage and the device isn't what I expected.
I wish I could have more hope that Android will ever decouple itself enough from Google to be useful/practical as free software. Unfortunately, it looks like it'll end up going the other way, and freedom-respecting Android ROMs will likely be niche and only supported on a narrow range of hardware.
I imagine the Linux phone has better long-term potential, as it's not tied to a proprietary ecosystem.
This makes me happy!
With LineageOS and microG in place a lot of work is already done in terms of having a free and open Android compatible mobile OS.
The FSF is now tackling the difficult part with firmware binaries that seems to stop a lot of phones from being supported.
This will take time, but as long as progress is made this will be awesome in the long run.
top 28 comments