Hi Alfonso! On 03/28/2017 09:55 AM, [email protected] wrote:
Astian OS is an operating system based on Firefox OS but unlike B2G, we want to continue with the original idea to continue with the maintenance of Gecko and where we avoid replacing it with Android.
There is some misunderstanding here. The idea with B2GOS is not to replace gecko by Android - it's too use the android widget layer of gecko since gonk is not supported anymore. Yes, that means that overall we would rely on a fatter stack.
One of our major goals is to bring Astian OS to more powerful devices, to implement new technologies such as WebGL and others to consolidate a competitive operating system, which is not only low-end but also high-end devices and desktop devices with web technology .
WebGL has been supported on b2g since day 1. Actually one of the first homescreens was written in WebGL by cjones. Git archeologists can likely find it in a repo somewhere!
We also believe in the convergence between devices. Astian OS is developed by the Astian Foundation, a non-profit organization that develops free software and technologies. Astian OS is our main project and in which we are investing all our time and energies.
I'm worried that you are wasting your time and energy on a project based on flawed assumptions. I already talked to you and others in the Astian telegram channel about that, but let me explain once again: A lot of b2g specific code is being removed from Gecko (see https://bugzilla.mozilla.org/show_bug.cgi?id=1306391). This means that you have the following choices: - Stay on an old gecko. I think that version 48 is the latest one with full b2g support. This exposes your users to potential security issues and that's quite irresponsible to base a new project on this version. There is no ESR version with security backports that you can rely on since the current ESR is based on gecko 52. - Move to gecko trunk, and backport the apis you still want to use. Given that you seem to plan keeping support for mozApps, this practically means forking gecko, and that practically impossible given the kind of changes involved. That is well understood and agreed on by people familiar with the codebase.
Unless you explain *how* you will solve these issues, it's hard to take Astian seriously.
We learned enough working on b2g to know how to design a better OS - getting stuck with the initial ideas as if they were some kind of sacred mantra is not the way to make progress.
We know that not all the choices we made in b2g were right. Some things didn't work (eg. the new app model didn't get traction), some were not the best technical choice (eg. telephony api exposed as DOM apis). That's why we spent some time last year to modernize and simplify the b2g codebase in order to limit the b2g-specific bits. We wanted to have as little coupling between the web runtime (gecko) and the UI layer, so we came up with this solution: - The "system" UI (current Gaia) will run as currently, using chrome:// urls. - Hardware related APIs that are not web standards will be exposed by external daemons. Only the system UI will be able to talk to these daemons, likely through a secured websocket connection. - We can't rely on Gonk being available, so we will start by relying on the Android widget and media support, but running a single app. Nothing is being drawn from usual Android.
So far progress has been very slow on b2gOS (thanks to vmx who is doing great work on the daemon side!) but fragmenting the community of people that still care about building a web tech based OS will not help either.
I think there is a need to take a cold, hard look at what would such a project look like and what it would be based on. Which rendering engine? which base layer? What runs as chrome vs. content? Which hardware support?
If Astian wants to contribute to a better computing platform for the people, by all means let's start by discussing how to best reach that goal. Your current track is a dead end I fear.
-- Fabrice Desré Emerging Technologies Mozilla Corporation _______________________________________________ dev-fxos mailing list [email protected] https://lists.mozilla.org/listinfo/dev-fxos
