02/12/2015, 10:38 PM
I was going to revive this: webOS on Raspberry Pi? - webOS Nation Forums which is worth a read, but I decided to start a new thread. Why?
Raspberry Pi 2 on sale now at $35 | Raspberry Pi
This new version is ARM7 @900MHz with 1Gb of Ram. I don't know how the Broadcom chipset compares to Qualcomm, but the speed is almost that of a Pre2, 2/3 that of a Pre3 and double the RAM of both.
Possibly, this will address the concerns about the device's power expressed in the previous thread. Additionally, this should be fully compatible with the earlier devices which have been greatly optimised since the first launch.
So what are the potential benefits of porting webOS to RaspberryPi? Some thoughts:
- Many are clamouring for Android on the Pi. There is a port, but Android is a stripped down OS for mobile and this isn't supported by Google or RaspberryPi. It runs so slowly that it is completely unusable. I'm not aware of any 'mobile-type' OS available for this computer and those seeking one might jump at the chance of webOS.
- A common use of the RaspberryPi is as a media player or jukebox. We have seen the success of LGwebOS on TV's. A slick touchscreen interface could be very popular and easily adapted by developers.
- RaspberryPi is intended for students, 'makers' and hackers. webOS is very open and hackable, so should be attractive to this group.
- Any OS developments and apps from the RPi community could be available to the main OpenwebOS / LuneOS development.
- To date, there is estimated to be 4.5million RPis (all types) sold. That approaches the numbers of LG webOS TVs sold and is a broad user base - most of whom are likely to be tinkerers / developers / software engineers.
- A port has already been attempted. It is here: Raspberry Pi - WebOS-Ports You can see that Qt 4.8 was the stumbling block.
- It seems that Qt5 has now been ported: RaspberryPi | Qt Wiki | Qt Project & Qt5 | Raspberry Pi
Arguably, the RaspberryPi is not an ideal device for a port - certainly no one on the Raspbery Pi forums has completed one, but adding a touchscreen display offers a tablet style device or project control panel for which a webOS UI would be ideally suited. If a working version was made available, RPi users might quickly discover the ease with which apps and the platform itself can be developed.
Before going further, I should point out that unless the process is as easy as assembling parts and running a machine that pops a working OS out of the other end, I don't have the skills, so it's very unlikely I will do this. My purpose here is to gather information and stimulate a bit of discussion that will hopefully inspire a skilled individual or group. I assume if it was that easy, someone will have put out a port over a weekend by now. Note that I'm also not suggesting that webOS ports should do this as they are still building the system on devices best suited to it.
So, if the prerequisites appear to exist, what might prevent a port?
- To work on existing devices, LunaSys manager had to be completely rewritten (LunaNext) to function on a layer of Android drivers. As previously noted, there are no Android drivers for the RPi and an attempted port is unusable. However, I think the drivers are a layer in a software stack. OpenwebOS will run on a PC and as Linux drivers clearly exist for the RPi, porting with the original LunaSys manager should work.
- If the above is true, will LuneOS work with LunaSys? Or is it only compatible with LunaNext? Will a port be restricted to the 'barebones' of the original open-sourced code? Perhaps the missing parts relate largely to telephony. While there are RPi telephony or wireless projects, these will use different H/W and different drivers, so this can be probably be ignored or added according to use case. About other LuneOS parts, I'm not sure. (Update) According to eblade below, LunaNext may work on RPi with Qt5.
Heres another list. My ignorant guess of what is required to make a port.
My assumption is that having set up an environment to build an OS, assembling pre-existing parts and compiling them with a script is fairly straight-forward if you know your stuff. With a new build, the problems arise in the code - finding and fixing bugs and compatibilities so that the build works.
How to Make a Computer Operating System (with Pictures) - wikiHow
Building and Installing Software Packages for Linux: Introduction
Create Your Very Own Operating System With Linux From Scratch [Linux]
So please correct my assumptions, tell me how easy this will be or how it will never work. Add your own knowledge so I can update this post and perhaps we will have solid enough information that someone will make an attempt!