Arm Development and Raspberry pi

[Barujin]

I have a suggestion for an alternative that would achieve a similar goal. Why not develope ROS for a mini x86 pc like this one: http://www.mini-box.com/Intel-D2500HN-Mini-ITX-Motherboard? Why does it need to be an ARM like the RasPi? I understand the interrest in ARM, but it seems rather extreme for a project that is simply trying to recreate Windows.

[SomeGuy]

I have a suggestion for an alternative that would achieve a similar goal. Why not develope ROS for a mini x86 pc like this one: http://www.mini-box.com/Intel-D2500HN-Mini-ITX-Motherboard? Why does it need to be an ARM like the RasPi? I understand the interrest in ARM, but it seems rather extreme for a project that is simply trying to recreate Windows.

Microsoft is currently positioning itself to try and be a big player on the ARM CPU. If they succeed like they did in the desktop market, 20 years from now Windows could have almost 100% of the market. So in a world where all ARM software development is for Microsoft Windows, then A Windows compatible OS on ARM could have a great advantage.

Of course, I personally hope Windows for ARM falls flat on it's face. And it already looks like it won't succeed any further than replacing their old Windows CE.

[livestrong2109]

by SomeGuy » 24 Jun 2012 21:14

Barujin wrote:
I have a suggestion for an alternative that would achieve a similar goal. Why not develope ROS for a mini x86 pc like this one: http://www.mini-box.com/Intel-D2500HN-M ... otherboard? Why does it need to be an ARM like the RasPi? I understand the interrest in ARM, but it seems rather extreme for a project that is simply trying to recreate Windows.

Microsoft is currently positioning itself to try and be a big player on the ARM CPU. If they succeed like they did in the desktop market, 20 years from now Windows could have almost 100% of the market. So in a world where all ARM software development is for Microsoft Windows, then A Windows compatible OS on ARM could have a great advantage.

Of course, I personally hope Windows for ARM falls flat on it's face. And it already looks like it won't succeed any further than replacing their old Windows CE.

With Monkey Boy in charge... The problem windows faces on ARM isnt something that can be fixed with a degree in marketing. The era of quick fix OS solutions is over. Andriod is an established OS that has thus far survived in court, and iOS is well embedded in the enterprise market. Where is MS going to take share from, all the bases are covered. Windows Phone 8 looks amazing and the NT kernel is a major plus, however without greater support from handset makers and developers there is no way they have a hope in hell. There just isn't any market left and the last thing Devs want to do is code for a 4th platform. What MS needs to do is buy out RIM and embed all the BBES services into MS Active Sync. If windows phones offered end to end encryption they might be able to win back the Enterprise market on the back of end to end security. The side effect of this though would be that Google would then gain the same tech via Google Sync as they are licensing it from MS to enable Push on windows phone and iPhone.

Personally the phone wars have just started things are really going to heat up in the next couple of months if MS can win over Corporate Power Users. But a RIM buyout is a must for them.

[SomeGuy]

I'd rather have a "phone" without a built in computer. My old desktop Unisonic 6434 telephone will still be operational and useful long after the next generation of smartphones are sitting in your landfills leaking toxic chemicals in to the water poisoning your children.

[Aeneas]

I actually think there is a huge market for "smartphones", as of yet untapped. But it is most likely true that it belongs to Android.

My reasoning is:

1. Android can get "dirt cheap" - Apple never will, that is why iOS will never be for "cheap" solutions.
2. Components tend to get cheaper over time. There IS a "lower limit" to price, but it can get reeeally low.
3. In countries like China, India or many countries in Africa, "the digital age" does not come with the computer: computers are large, "whiny" and not connected. A cellphone, on the other hand, is cheap, small (transportable), reliable, and "connected over the air". Even if you do not have internet, it can send and receive SMS, and that is surely enough to tell your fishermen companions in your village what price to expect in the market today (it IS actually seeing such use, I recently read about it). Now, if you add to that some "computer capabilities" - e.g., taking notes, and some calculator, you already have an EXTREMELY useable device.
4. A smartphone, as a "small computer", could build on that easily.
5. Even in "civilised" countries, given that most people prefer "cool phones" (even though such people are mostly ignorant of the significant risk of state supervision, infection etc.), there will be a high demand and the scaling effects will work entirely to the advantage of smartphones.

I just see no real usage of ReactOS as of yet, but the effort may at least clean up the code so far and maybe prepare generally for ARM.

See, Android and iOS simply have hundreds of thousands of apps. THAT is their TRUE value; it is just like with Windows: no-one cares about the system, everyone just cares about the apps. How many does ReactOS have?

[DOSGuy]

See, Android and iOS simply have hundreds of thousands of apps. THAT is their TRUE value; it is just like with Windows: no-one cares about the system, everyone just cares about the apps. How many does ReactOS have?

However many Windows does. Is this a trick question?

[Aeneas]

Windows ON ARM?

[DeadDude]

Yep, Windows on ARM CPUs... heh...

I got a copy of QEMU for iOS with an image of Win95 linked to it. It boots in about 15 minutes, but does actually boot.

My ipod touch is the original model, so I reckon newer devices would be quicker...


Point of me posting this- it is apparently completely possible, so why not?

Just see if you can find an emulator for your device, and remember that it is going to be slooooooooooooooow.

Why reinvent the wheel a second time, when there could be decent emulators in the first place to run ROS on top of?

[EmuandCo]

Emulators will always be inefficient. There is windows rt in the future. There was windows ce in the past. There always was and will be windows for arm

[Aeneas]

I understand your point, but simply put: I see talk, and talk only - I see nowhere 0.3.15. Neither do I see 0.4.0.

Your efforts are undoubtedly intense and successful, but you ARE 30 people, and this OS does not yet even work on its "primary target". Linux in 2001 was relatively way more useable than ReactOS is now. That applies IMHO for the BSDs, too, and they get but a tiny fraction of the funding of Linux...

I would say, concentrate - it is pretty much your only chance of becoming alike to BSD and not to FreeDOS. But of course, not being a developer, I have no say in this whatsoever.

[Haos]

You do realise that in project alike this one, you cannot force anyone to do anything bar he wants to do at the given time?

[Aeneas]

Not by force, but by attractiveness of the functionality and the strength and quality of the organisation do you achieve success.

I mean... look at BSD, as I said. (Linux I do not mention, as the media hype gives it a disproportionate advantage.) You cannot tell me that BSD is "such" a strong force. And yet they do get a USB stack, they improve their filesystems, even the tiny NetBSD is now making a new (modular) kernel design.

"Somehow" they do it. ReactOS certainly has possibilities, too; you are in a German association, a "Verein", after all. But to go around telling me how wonderful it is to start ARM development - sorry, looking at the x86 stage, I cannot but find this laughable! It is as if I made a cute and nice paper airplane and go around concluding contracts for military jets...

[zydon]

I think Linux and BSD couldn't cut a mustard for Windows users. No point mentioning to us repeatedly because we were majority OS users want Win32 based OS. Developers time is not the core problem ROS take so long to evolve. To find the different method doing the same thing as Windows OS is where lies the problem.

No other open source Windows OS Clone out there has achieved the level of ROS to this far. If those the same Linux and BSD teams could make a Windows 9x clone at the same speed as they did on Linux and BSD, then people can complain about ROS project. Otherwise, try to think of anything useful you could contribute to this project so it could progress more and you get improved ROS for a free.

Building the ARM version of ROS would probably helps solving some of those issues when ARM devices developers and hobbyist could get an open source Win32 based OS available for the processor. Definitely there where tough decision to be made whether to provide a VM like like the Android OS to run existing PE applications or using separate version of PE for ARM processors. Based from current scenario, software developers attracted to develop applications for the OS which is can run on both IA32 and ARM processors.

So, when ROS for ARM able to boot into it's own x86 VM, developers who work on ARM processors will take it seriously and the progress will become much more faster.

[BrentNewland]

I have a suggestion for an alternative that would achieve a similar goal. Why not develope ROS for a mini x86 pc like this one: http://www.mini-box.com/Intel-D2500HN-Mini-ITX-Motherboard? Why does it need to be an ARM like the RasPi? I understand the interrest in ARM, but it seems rather extreme for a project that is simply trying to recreate Windows.

The point is not to uild ReactOS on a small board, the point is to make it easier to port ReactOS to ARM processors. That and the Raspberry Pi has an ARM processor that is (IIRC) the first one that makes it possible to port ReactOS, due to changes in the memory manager (or something like that). Starting with that as a baseline should make it much easier to support later ARM processors.

There are already ARM motherboards with PCI-Express slots. It's going to become more common.

More specifically though, ReactOS has the potential to provide better multitasking on a phone than any other mobile OS. How many mobile operating systems allow you to quickly see all your active/running programs? Allow you to easily browse all files on the system? Provide fast switching between programs? Make it easy to close programs? ReactOS could take care of all that.

So, when ROS for ARM able to boot into it's own x86 VM, developers who work on ARM processors will take it seriously and the progress will become much more faster.

Booting ReactOS into a virual machine on an ARM processor does nothing to make it easier to port ReactOS to ARM, nor does it make it more desirable.

However, being able to run x86 Win32 programs on an ARM processor (even at a reduced speed) would certainly be met with great interest.

you are in a German association, a "Verein", after all. But to go around telling me how wonderful it is to start ARM development - sorry, looking at the x86 stage, I cannot but find this laughable! It is as if I made a cute and nice paper airplane and go around concluding contracts for military jets...

ReactOS Foundation is based in Russia, not Germany, I believe.

It's great if you like to laugh at ReactOS, but doing so on the ReactOS forums is the same as trolling.

And ReactOS is actually in pretty damn good shape right now.

I understand your point, but simply put: I see talk, and talk only - I see nowhere 0.3.15. Neither do I see 0.4.0.

Your efforts are undoubtedly intense and successful, but you ARE 30 people, and this OS does not yet even work on its "primary target". Linux in 2001 was relatively way more useable than ReactOS is now. That applies IMHO for the BSDs, too, and they get but a tiny fraction of the funding of Linux...

I would say, concentrate - it is pretty much your only chance of becoming alike to BSD and not to FreeDOS. But of course, not being a developer, I have no say in this whatsoever.

As the developers have said, numbers don't matter, only progress. And ReactOS is making great progress. The number of "major" features to implement is steadily dropping, and I believe development is at (or approaching) the stage where focus is mainly on finding bugs that are preventing programs from running (and fixing them), and implementing smaller unimplemented features.

ReactOS has had a lot less support than Linux. Linux has had tens (if not hundreds) of thousands of contributors, and the benefit of deciding everything from scratch (not reverse engineering a closed system). ReactOS is more usable than many Linux distros in 2001.

As far as "concentrating", it's important to get big items like this out of the way as soon as possible. If they could get the ARM port done this year and it added a year to the completion of ReactOS, that would be much better than adding several years after the completion of ReactOS for the ARM port.

[Aeneas]

"ReactOS has had a lot less support than Linux." - But that is not an "excuse", that rather belongs to the problem sphere. Actually, when doing a working clean room re-implementation of Windows, people should be kicking in your doors from wishing to participate... If progress matters, that should matter IMHO, too.

Actions speak louder than words. Let me say it this way: I hope that you will be perfectly right, that progress is great and results, making this progress tangible in terms of increased usability, will follow soon. We shall see what of such hopes materialises into reality. We have the year 2012 now. WinXP's support runs out in 2014. This is, in my eyes, the time window that is still open. (Everything beyond becomes either unrealistic or is to be counted to "FreeDOS-type" development, i.e. something that will arrive way too late to be actually influential.) But within the time window, ReactOS is going to have the advantage of not needing to take care of applications - these are already implemented all over the world...

(And to make it clear: ReactOS is actually one of my total "pet projects", because it is a NON-Unix one. I really do like the Unixoids, particularly the BSDs, but their trend to completely overtake everything, from washing machines to smartphones to servers to supercomputers, appears to me somewhat threatening to OS-diversity. I am actually quite sorry that OpenVMS and OS/2 are basically dead.)