Project

General

Profile

Replicant's big moment.

Added by Chris _ over 7 years ago

-edit I just realized I mistakenly posted this in the development forum, was meant for the discussion one, my apologies.

Hi, I've recently dipped my toes into the world of free software and as a new member of the wider free software community I wonder if I can bring some fresh eyes to this project. A little background on me, I got my interest in open-source after using Android for the last few years, using the F-Droid repository & it's wealth of great applications that it provides. I've always been a Windows user, but I recently picked up an ASUS Flip c100p, one of the first of three Chromebooks that have the ability to run Android apps with the help of a (currently closed-source) Google-developed container for Android. I've since tested 100+ apps from the F-Droid repository using this device & have even started toying with linux using chroots via a project called crouton (which is maintained by a Googler on Github). Basically, I'm convinced that Android is about to take off in a major way (even more than it already has) because the applications work so well even in the completely optimized state they're currently in. More importantly, Google is deprecating Chrome Apps for all platforms other than ChromeOS, and is pushing devs to make Android apps with good mouse/keyboard support as well as support for multiwindow. Soon, Google is going to require that all Chromebooks manufactured include GPS/accelerometers/fingerprint-reader/etc as all future Chromebooks will support this new container for Android. They're about to make a huge push for Android to be so much more than just a mobileOS & with the right vision it could revolutionize the dream of mainstream adoption of Linux on the desktop if that container-tech source code is made available or if a team tries to develop one on their own. Something like that combined with a distro like Papyros could provide a cohesive GNU/Linux + Android OS, and given that Android is already used by literally billions of people around it'd make switching over so much easier than it is today given that people are already familiar with Android's GUI.

tl;dr

I understand that the goal of Replicant is to provide people with a truly libre mobile-OS solution, but given that Android is moving toward becoming an operating system for both mobile and desktop environments, the project leaders here should consider expanding their vision. From what I understand, this project is mostly maintained by a fairly slim group of devs, by supporting Android more broadly it would make waves in the community, create headlines, and potentially draw in more devs. As a result, support/maintenance of older devices could become easier & the addition of new devices would become less daunting.

Here's what I recommend considering in order to achieve this goal:

- I'll start with something easy(?), the Replicant SDK is capable of creating virtual devices and from what I can gather its being updated to 6.0. There exists a project called Shashlik (which is free/open-source) which leverages that same technology, strips it down, and modifies it to create an Android Emulation Layer to easily run apps inside of Linux. From my (admittedly novice) understanding of coding, it's easy to imagine simply compiling a version of Shasklik that substitutes the Google-provided SDK for the one produced by the Replicant project. When used on top of hardware like one of the FSF endorsed Thinkpads, it would be the first straightforward implantation of Replicant on fully free hardware. Most importantly it'd give everyone with an x86 processor the ability to test drive a fully free version of Android.

- Make supporting the EOMA68 Libre Tea Computer Card a priority (runs an Allwinner A20 Dual-core ARM Cortex A7 processor), as it appears the project will be endorsed by the FSF. The project leader seems to be very friendly & is very open to working with devs, it's not unimaginable that he'd be willing to send out some prototypes (I'd even personally be willing to donate some $$ to help subsidize the cost for any interested devs). I know nothing about the process of adding support for a device, but given that it's ARM hardware, it seems to a great candidate for porting, and since the hardware is fully free, it seems that it might(?) be possible to make this the first time that Replicant could leverage hardware acceleration for graphics, a massive milestone in my eyes. Most importantly, it'd be the first ever libre piece of hardware to be naively supported by a libre fork of Android. Imagine the headlines if this is achieved, and imagine the interest/new members it could draw to the project!

- Stretch goal, add support for Android x86. I've read that this one's been shot down in the past, but if either one of the above goals could be achieved, it's not inconceivable that with greater interest in the project that an interested dev could take up the challenge.

In closing, I hope that the ideas I've shared here don't come across as demanding/pushy my only intention is to share what I find to be the most fascinating development in the world of free/open-source and illustrate some tangible ways in which Replicant could be a part of it. I think that the future looks brighter than ever before for Android and GNU/Linux generally and with the work you do here you're helping to ensure that future can be free. Big thanks to all the devs here for their continual work on the project.

Sources:

Ars - "The Play Store comes to Chrome OS, but not the way we were expecting" (Interview with the devs of the Android container software for ChromeOS, doesn't rule out the possibility of the tech being open-sourced)
http://arstechnica.com/gadgets/2016/05/the-play-store-comes-to-chrome-os-but-not-the-way-we-were-expecting/

Papyros (Material-designed Linux distro, unfortunately it appears that development has stalled)
http://papyros.io/

Wired - "How Chromebooks Are About to Totally Transform Laptop Design" (explains Google's vision for the future of ChromeOS/Android)
https://www.wired.com/2016/09/chromebooks-totally-transform-laptop-design/

Shashlik (Android Emulation Layer project)
http://www.shashlik.io/

List of hardware currently endorsed by the FSF
https://www.fsf.org/resources/hw/endorsement/respects-your-freedom

Crowdfunding page for the (fully-funded) EOMA68 ARM-based computer (soon to be endorsed by the FSF), and a video interview with it's creator (in it, he talks about his openess to working with devs)
https://www.crowdsupply.com/eoma68/micro-desktop

https://www.youtube.com/watch?v=GspUDdNlsng


Replies (1)

RE: Replicant's big moment. - Added by Wolfgang Wiedmeyer almost 7 years ago

Thank you for providing your thoughts on this and documenting related info! I don't think that there is general opposition from Replicant to add support for more desktop-like or x86 platforms. There is just nobody working on it right now.

I'm not familiar with the different container and emulator technologies for Android, so I can't comment on the technical aspects. Is Shashlik really using the Google SDK as is? By glancing over the source code repos, this doesn't seem to be the case. They seem to be using a minimal set of the frameworks repos and modifying them.

I'm not sure if modifications or added free implementations that are done by Replicant are even relevant for these projects. If the concern is that they are not fully free or advertise non-free software, then it's probably the easiest to fork these projects and remove/replace the non-free parts. Whether the fork should then be its own project or could be hosted under the umbrella of the Replicant project needs to be discussed for specific cases. It doesn't always make sense to merge projects, even if there is some common source code and common goals exist.

If a different project is already a lot closer to run Android applications on FSF-endorsed hardware by only using free software, then this is great and people should contribute to these projects! There is no need to integrate or fork these projects, only to rubber-stamp them with the Replicant name. But if it really makes sense to join forces and if there are people willing to work on this on the Replicant side, then this should be done.

If someone wants to port Replicant to the EOMA68, go for it! I'm already busy with other hardware I want to get working. And I think Paul and Denis are as well.

As for the Android-x86 project, it's basically the same. The quickest way is to fork it and remove the non-free parts so that Android-x86 images only contain free software. The development work for this would be pretty much the same as the initial work to get it merged with Replicant.

    (1-1/1)