3 ms·

The decision to develop the first GNOME phone while also developing a hacker-friendly privacy-respecting physical device is going to be the death of this projec

The decision to develop the first GNOME phone while also developing a hacker-friendly privacy-respecting physical device is going to be the death of this project. I don't understand why they didn't just focus on the hardware, make it very hackable for the community to develop the GNOME (or any other) OS for it. Isolate the baseband from the application processor and all the other privacy-respecting hardware decisions, land all the changes required for a mainline kernel to support everything upstream, and release it with AOSP running on it. Then invest in the community to replace AOSP (and sell more of your hardware to said community as it grows).

craftyguy 8y ago
> I don't understand why they didn't just focus on the hardware, make it very hackable for the community to develop the GNOME (or any other) OS for it. Probably because the "do everything yourself hackerz" are not the only target audience.. This company has a history of selling quality laptops with their PureOS distro on it, while also supporting the ability to put on other distros for those that want to.
kop316 8y ago
I purchased both an OpenPandora (https://pyra-handheld.com/boards/ https://pyra-handheld.com/boards/) and a Novena (https://www.kosagi.com/w/index.php?title=Novena_Main_Page https://www.kosagi.com/w/index.php?title=Novena_Main_Page). The OpenPandora came out with a fairly good OS to start out with, and they have continually supported it, so it has still been useful even 7 years after it was released. If it wasn't for the fact that I can no longer find it, I have no doubt that I would still be using it. I am a supporter for the Dragonbox Pyra, and I am hoping it will have the same support. I cannot say that I regret buying my Novena (as I learned a lot from it), but I rarely use it. I have to manually update the kernel and bootloader myself, and it is stuck on Linux 4.4, as the support and community just aren't there to help. I was able to figure out how to update it to the newest Debian, but there is no official support for it, so I imagine most folks are still on the previous version. As it stands, it marginally functions, but if I need to use a laptop again heavily, I will simply just buy a new laptop. I relate these two stories because I think the Librem 5 will need to show that it has a polished software stack day one to gain traction and grow a community. The Open Pandora was successful in that, and that allowed the creator to make two iterations of the Open Pandora and create the Pyra. The Novena did not have that nearly has much, and as a result, I am very doubtful that there will be a Novena + 1, and frankly after seeing the support it got, I am hesitant to purchase any other products made by Bunnie.
- newnewpdro 8y ago
  None of the devices you mentioned were polished physical products competitive with their contemporaries, while also costing significantly more. What the community needs is a compelling physical piece of hardware with first-class upstream driver support for everything and relatively modern specs. We can iterate on the userspace as a community. But there won't be a growing community if you don't sell us a compelling open piece of hardware to run it on. We can't easily iterate on the hardware as a community, this is the part puri.sm should focus on delivering. Sell us a modern device that respects our privacy and ability to replace boot loaders and run mainline kernels. Give use first-class driver support for everything included from the GPU to the cellular modem and cameras and all the APIs necessary to speak with them. It just seems like a total disconnect from the reality of mainline kernel support for smartphone hardware, to be off in la-la land working on a GNOME phone OS before delivery. Pavel Machek's talk from 10/22/2017 [1] comes to mind as representative of the state things are in, in mainline. I'm far more confident in the GNOME/free software community's ability to write dialers, messengers, email and photo apps than its ability to quickly bring up low-level support for such hardware and develop the new kernel APIs where things like V4L2 fall short. [1] https://lkml.org/lkml/2017/10/21/181 https://lkml.org/lkml/2017/10/21/181
  - kop316 8y ago
    "None of the devices you mentioned were polished physical products competitive with their contemporaries, while also costing significantly more." I would actually argue the Open Pandora was a polished physical product. It worked incredibly well when it just came out. "What the community needs is a compelling physical piece of hardware with first-class upstream driver support for everything and relatively modern specs." Correct me if I am wrong, but aren't the Librem 5 folks trying to do just that? They chose an imx8 because it has really good upstream support, and then they can add upstream support for the peripherals.
curt15 8y ago
>I don't understand why they didn't just focus on the hardware, make it very hackable for the community to develop the GNOME (or any other) OS for it. This. Why do they insist on a mobile GNOME interface from scratch instead of building on more mature mobile UI frameworks like Qt5/Plasma mobile? Are there some deep philosophical differences? Qt already has a lot of experience with touch interfaces.
- linuxftw 8y ago
  Their area of expertise is making a pretty linux distro for their custom laptop line, that's why. I agree with you and GP. Every dollar invested into gnome is a waste IMO. Make the hardware, make it easy to build an image from source for it, and then focus on the gnome portion.

craftyguy 8y ago

> I don't understand why they didn't just focus on the hardware, make it very hackable for the community to develop the GNOME (or any other) OS for it. Probably because the "do everything yourself hackerz" are not the only target audience.. This company has a history of selling quality laptops with their PureOS distro on it, while also supporting the ability to put on other distros for those that want to.

kop316 8y ago

I purchased both an OpenPandora (https://pyra-handheld.com/boards/ https://pyra-handheld.com/boards/) and a Novena (https://www.kosagi.com/w/index.php?title=Novena_Main_Page https://www.kosagi.com/w/index.php?title=Novena_Main_Page). The OpenPandora came out with a fairly good OS to start out with, and they have continually supported it, so it has still been useful even 7 years after it was released. If it wasn't for the fact that I can no longer find it, I have no doubt that I would still be using it. I am a supporter for the Dragonbox Pyra, and I am hoping it will have the same support. I cannot say that I regret buying my Novena (as I learned a lot from it), but I rarely use it. I have to manually update the kernel and bootloader myself, and it is stuck on Linux 4.4, as the support and community just aren't there to help. I was able to figure out how to update it to the newest Debian, but there is no official support for it, so I imagine most folks are still on the previous version. As it stands, it marginally functions, but if I need to use a laptop again heavily, I will simply just buy a new laptop. I relate these two stories because I think the Librem 5 will need to show that it has a polished software stack day one to gain traction and grow a community. The Open Pandora was successful in that, and that allowed the creator to make two iterations of the Open Pandora and create the Pyra. The Novena did not have that nearly has much, and as a result, I am very doubtful that there will be a Novena + 1, and frankly after seeing the support it got, I am hesitant to purchase any other products made by Bunnie.

newnewpdro 8y ago
None of the devices you mentioned were polished physical products competitive with their contemporaries, while also costing significantly more. What the community needs is a compelling physical piece of hardware with first-class upstream driver support for everything and relatively modern specs. We can iterate on the userspace as a community. But there won't be a growing community if you don't sell us a compelling open piece of hardware to run it on. We can't easily iterate on the hardware as a community, this is the part puri.sm should focus on delivering. Sell us a modern device that respects our privacy and ability to replace boot loaders and run mainline kernels. Give use first-class driver support for everything included from the GPU to the cellular modem and cameras and all the APIs necessary to speak with them. It just seems like a total disconnect from the reality of mainline kernel support for smartphone hardware, to be off in la-la land working on a GNOME phone OS before delivery. Pavel Machek's talk from 10/22/2017 [1] comes to mind as representative of the state things are in, in mainline. I'm far more confident in the GNOME/free software community's ability to write dialers, messengers, email and photo apps than its ability to quickly bring up low-level support for such hardware and develop the new kernel APIs where things like V4L2 fall short. [1] https://lkml.org/lkml/2017/10/21/181 https://lkml.org/lkml/2017/10/21/181
- kop316 8y ago
  "None of the devices you mentioned were polished physical products competitive with their contemporaries, while also costing significantly more." I would actually argue the Open Pandora was a polished physical product. It worked incredibly well when it just came out. "What the community needs is a compelling physical piece of hardware with first-class upstream driver support for everything and relatively modern specs." Correct me if I am wrong, but aren't the Librem 5 folks trying to do just that? They chose an imx8 because it has really good upstream support, and then they can add upstream support for the peripherals.

curt15 8y ago

>I don't understand why they didn't just focus on the hardware, make it very hackable for the community to develop the GNOME (or any other) OS for it. This. Why do they insist on a mobile GNOME interface from scratch instead of building on more mature mobile UI frameworks like Qt5/Plasma mobile? Are there some deep philosophical differences? Qt already has a lot of experience with touch interfaces.

linuxftw 8y ago
Their area of expertise is making a pretty linux distro for their custom laptop line, that's why. I agree with you and GP. Every dollar invested into gnome is a waste IMO. Make the hardware, make it easy to build an image from source for it, and then focus on the gnome portion.