SoC Poll - What's most important for you?

[CrazyIvan]

If the Pyra would run Windows, I think a big part of the current community (and most of the devs) would immediately say "meh" and lose interest.

ED said many times that pyra would run Linux. It may or may not run any number of other OSes. If multiple OSes would be supported then Linux would be in that list. He's not Craig you know.

Pyra won't be the only portable linux machine but it'll still be the portable linux machine with hardware keyboard and gaming controls. If a community member don't want it than probably it'll lose interest anyway and move to another device be it a powerful gaming handheld, linux tablet or phone or an arm netbook. I believe pyra won't be as special as pandora when it'll be out

 

[_wb_]

If the Pyra would run Windows, I think a big part of the current community (and most of the devs) would immediately say "meh" and lose interest.

ED said many times that pyra would run Linux. It will or will not run any number of other OSes. If multiple OSes would be supported then Linux would be in that list. He's not Craig you know.

Pyra won't be the only portable linux machine but it'll still be the portable linux machine with hardware keyboard and gaming controls. If a community member don't want it than probably it'll lose interest anyway and move to another device be it a powerful gaming handheld, linux tablet or phone or an arm netbook. I believe pyra won't be as special as pandora when it'll be out

I didn't say "if it would only run Windows", I think that even if it would run both Windows and GNU/Linux (e.g. ship with a dual boot), many of us would be less interested. Because that would mean that the community would be fragmented, and the audience for Pyra-specific GNU/Linux applications would be even smaller than it would already be otherwise.

I hate Windows that much that I think it's a useful feature when a particular SoC does not run Windows

 

[hmc]

_wb_, just because a device can run Windows doesn't make the device a Windows device. You can still run other operating systems on it.

And I don't mean an x86 Pyra should be tied to Windows! By all means - NO! (I also hate Windows).

I meant to offer it as an option only. Just maintain compatiblity in terms of boot manager and hardware driver availability (and what might be necessary in addition).

CrazyIvan: I agree that it should not be our goal to attract entire companies to equip all of their employees with Pyras. I meant more that kind of people who work for a company that uses Windows / Exchange etc. (so you need a Windows device to integrate with your company), and that has the "bring your own device" policy. There will be people, more or less geeky ones, that prefer to bring a Pyra rather than a laptop, tablet or convertable. Or maybe as an alternative for a Blackberry or similar ultramobile device. That new market won't lever the Pyra out of its market niche. It would only make that niche a little larger. We don't speak about huge sums of money here, and OpenPandora GmbH should not be the main supplier of IT hardware for big companies. That was not what I intended with my suggestion.

And if this group of people using a Pyra for business stays small enough, of course I can offer support for them, too. 

 

[commander-beef]

I agree with hmc..

x86 should be the only way to go for EDs company

BTW. In mid nineties i've seen a lot of atari portfolio in delta co. company in Poland, so if there was a a demand in mid 90's for mini computer , why there could not be demand today. We as a community of pc enthusiast know already that tablets failed miserably in todays business 

 

[_wb_]

Also note that if you want it to legally run Windows, you probably have to sign some kind of exclusivity contract with Microsoft that disallows you from shipping or supporting any other OS. Of course we can still all install and unofficially support GNU/Linux, but not without paying for Windows first. At least that's how I understand things.

 

[_wb_]

I agree with hmc..

x86 should be the only way to go for EDs company

Why? For the closed-source x86 software? If you like closed-source software that much, what about the closed-source ARM software? Don't you think that closed-source ARM software will tend to be more suitable for a handheld device, in terms of power consumption, GPU expectations, physical screen size expectations, etc?

Anyway, I don't think we should let the architecture be dictated by considerations about closed-source software compatibility. Many of us aren't even interested in closed-source software.

 

[SeeThruHead]

The point is options. x86 can run android linux windows, whatever you want basically, as long as there are drivers. Giving people more choice is never a bad thing. You don't necessarily have to develop for or actively support those choices but the fact that they aren't taken from the end user for no reason is a good thing. All your open software will run on x86 so I don't see the problem.

Another arm handheld right now is just a pandora with a faster SoC and better screen. Now faster is always nice but how is the pyra going to use that power? Will it enable users to run software that they could not before? Or just run the software that currently already runs on the pandora but with longer battery life? I think it would be more beneficial if the choice of SoC allows for the greatest variety of software to be run on the device, whether that is applications or os's. Choice is king. (Of course that all hinges on whether the x86 soc is sourceable and safe in a business sense but that is for ed to decide.

 

[commander-beef]

I agree with hmc..

x86 should be the only way to go for EDs company

Why? For the closed-source x86 software? If you like closed-source software that much, what about the closed-source ARM software? Don't you think that closed-source ARM software will tend to be more suitable for a handheld device, in terms of power consumption, GPU expectations, physical screen size expectations, etc?

Anyway, I don't think we should let the architecture be dictated by considerations about closed-source software compatibility. Many of us aren't even interested in closed-source software.

do i said somewhere i want closed software? Nope. 

Linux with x86 pyra should be good choice

 

[_wb_]

A faster SoC, whether it is ARM or x86, with access to a big GNU/Linux distro like Debian, means a lot of software will become available that was not really usable on the Pandora because it would be too slow, it would need a higher screen resolution, or just nobody made the effort to port and package it as a PND. There are over 37500 software packages in Debian.

Some of the new things I imagine you will be able to do on the Pyra but not (really) on the Pandora:

- view any kind of movie format effortlessly

- simple movie editing

- advanced image manipulation

- 3D modelling, raytracing etc

- scientific software with decent performance (still no super computer number crunching of course)

- programming in any programming language, however obscure

- faster and more reliable web browsing and other internet stuff

- if it has a GPS: map viewers, route trackers, navigation etc

- office applications like LibreOffice will be more comfortable with a larger and higher res screen

- PSP and Dreamcast emulation full-speed

- all previous systems emulated with way less battery consumption and/or nicer scale filters

- maybe emulation of a few more recent systems, at some point

I don't care about x86 or ARM. GNU/Linux and Android work on both (Android probably better on ARM and closed-source linux stuff better on x86). I don't want to run Windows, and I don't think that even if the SoC is x86, the Pyra will be able to run Windows -- who is going to convince Microsoft to make it work? It's not just a normal PC, there is no way that a normal Windows will just install out of the box. And even if Microsoft for whatever reason decides to support the Pyra, there's no way they will do that without demanding exclusivity.

So you can talk about options and choices all day, but the reality is that whatever SoC you pick, some options are eliminated, at least when you're interested in closed-source software. An x86 SoC might give some extra options for running some windows software through Wine and some closed-source linux games, but it will take away the option to run some Android software and ARM-optimized software (open or closed source) like most of the emulators we have on the Pandora. It is by no means clear to me which SoC "allows for the greatest variety of software to be run on the device". Also note that we cannot predict future closed-source software releases, but with the OUYA and the Nvidia Shield, along with all those smartphones and tablets out there, it could very well be the case that ARM will be a bigger target than x86 in the coming years.

I agree with hmc..

x86 should be the only way to go for EDs company

Why? For the closed-source x86 software? If you like closed-source software that much, what about the closed-source ARM software? Don't you think that closed-source ARM software will tend to be more suitable for a handheld device, in terms of power consumption, GPU expectations, physical screen size expectations, etc?

Anyway, I don't think we should let the architecture be dictated by considerations about closed-source software compatibility. Many of us aren't even interested in closed-source software.

do i said somewhere i want closed software? Nope. 

Linux with x86 pyra should be good choice

I didn't say it's not a good choice, but you explicitly said that x86 is "the only way", so please explain to me why you said that, if it isn't for closed-source software.

 

[Hồng Thất Công]

I now lose interest in Pyra after reading this thread

 

[EvilDragon]

I now lose interest in Pyra after reading this thread

Me too.

That's it folks, good bye!

 

[Foxgod]

I think the votes have spoken anyway, people want an open source handheld with arm and linux.

 

[WizardStan]

Me too.

That's it folks, good bye!

That's understandable.

 

[hmc]

I agree with hmc..

x86 should be the only way to go for EDs company

That's not what I said!

I said, that my prefered way is FIRST ARM, THEN x86.

The point is options. x86 can run android linux windows, whatever you want basically, as long as there are drivers. Giving people more choice is never a bad thing. 

That's not quite right. The more choices you have, the more frustrated you can become, because making a good and well-thought descision is overwhelming your brain. There are a lot of recent publications about this phenomenon. And I can confirm this! 

Hence I previously agreed to this: ;-)

 Let ED and his team make the decision behind closed door. 

 

 

Me too.

That's it folks, good bye!

oops. Not good...

 

[EvilDragon]

Me too.

That's it folks, good bye!

oops. Not good...

Haha, anyone taking me seriously here?

Nope, not a chance

As voting looks like, people are more interested in ARM, which is also my idea.

While x86 looks interesting sometime in the future, we don't know enough about it yet, and I rather have a small stable production run working before thinking about going big with an x86 mini PC and huge sales.

It wouldn't be the first company that has died because it grew too strong.

We also don't have development kits available. I'll contact Intel and ask about that, but I'm not sure they care about us... and if they had a development PCB, they'd announce it somewhere, I guess.

Also, the reason I started my shop and everything was because I like niches

Anyone who lost interest because of this thread will probably regain it once we've got more videos showing of what it can do

 

[hmc]

Not really taking you serious, since you have started this discussion with the poll

Did you take me serious? ;-)

The Pyra will certainly stay a niche product. If not for architecture, at least the gaming controls will prevent that large companies equip their staff with Pyras. ;-) It already looks like a geeky device. So it must be one. 

Sorry if I scared some people with terms such as "business customer" and "Windows", "Outlook" etc. - To clarify it once again: I do NOT, NEVER aim at having a Windows-only Pyra for the mass market. 

I also like niches. See my website as a proof ;-)

I was myself an employee of a Windows-based company for many years, and I would have loved to have an x68-based Mini PC with the option to run Windows on it, just to be able to also use it for work. I tried exactly this with the Viliv N5 and UMID M1, but failed, because those devices did not have the quality and flexibility I would have needed.

Btw: Now I don't have anything to do with Windows anymore, not even at work, but with OS X and Linux instead. *HAPPY*

 

[Neelix]

I don't care about x86 or ARM. GNU/Linux and Android work on both (Android probably better on ARM and closed-source linux stuff better on x86). I don't want to run Windows, and I don't think that even if the SoC is x86, the Pyra will be able to run Windows -- who is going to convince Microsoft to make it work? It's not just a normal PC, there is no way that a normal Windows will just install out of the box. And even if Microsoft for whatever reason decides to support the Pyra, there's no way they will do that without demanding exclusivity.

Some time ago WizardStan made a statement along the lines of that it would be impossible to install MS DOS on a modern PC.  As it happened I was able to prove that to be incorrect because I had (on a lark) already successfully installed MS DOS 6.22 on my Atom based netbook. (booting from a USB key no less, and playing Kings Quest with a USB mouse, hardware abstraction is a wonderful thing)     If there was an x86 version of the Pyra, I've no doubt that someone would find a way to make it boot windows.  It might not work very well, and MS would (most likely) never support it,  but I would be surprised if it were actually impossible.

As a friend of mine likes to say,  Nothing is impossible,  though probability may vary.

- Neelix

 

[_wb_]

61% want to stay with ARM, and 19% don't care, so that is 80% who like ARM versus only 39% who like x86. So if you want to follow this poll, it should be one of the three ARM SoCs.

The overwhelming majority is concerned about battery life (only 8% does not care about it, and 32% would sacrifice power to get more battery life), and for a majority this is also the top priority. So I think we need to get information about the (real or expected) perf/W curves for those SoCs, and get the SoC that looks best.

18% does not care about FOSS at all, but over 80% do care about it. Most of them are pragmatic about it (only 8% take a principled position), and don't mind some blobs as long as it does not cause trouble (like being stuck with an old kernel or a buggy driver). The question is: do we want another PowerVR GPU? Are we sure we will not get ourselves into trouble with it?

If we want ARM and a GPU that has FOSS drivers, then Snapdragon is the only option. If its perf/W curves are better than, or close enough to those of the other two SoCs, I would go for Snapdragon.

 

[_wb_]

I don't care about x86 or ARM. GNU/Linux and Android work on both (Android probably better on ARM and closed-source linux stuff better on x86). I don't want to run Windows, and I don't think that even if the SoC is x86, the Pyra will be able to run Windows -- who is going to convince Microsoft to make it work? It's not just a normal PC, there is no way that a normal Windows will just install out of the box. And even if Microsoft for whatever reason decides to support the Pyra, there's no way they will do that without demanding exclusivity.

Some time ago WizardStan made a statement along the lines of that it would be impossible to install MS DOS on a modern PC.  As it happened I was able to prove that to be incorrect because I had (on a lark) already successfully installed MS DOS 6.22 on my Atom based netbook. (booting from a USB key no less, and playing Kings Quest with a USB mouse, hardware abstraction is a wonderful thing)     If there was an x86 version of the Pyra, I've no doubt that someone would find a way to make it boot windows.  It might not work very well, and MS would (most likely) never support it,  but I would be surprised if it were actually impossible.

As a friend of mine likes to say,  Nothing is impossible,  though probability may vary.

Yes, it would probably be possible somehow, if necessary by patching some executables etc. But it would most likely not be legal, and not officially be supported by either Microsoft or ED.

 

[EvilDragon]

18% does not care about FOSS at all, but over 80% do care about it. Most of them are pragmatic about it (only 8% take a principled position), and don't mind some blobs as long as it does not cause trouble (like being stuck with an old kernel or a buggy driver). The question is: do we want another PowerVR GPU? Are we sure we will not get ourselves into trouble with it?

Yes and no.

Yes: We never know how buggy any binary blob driver will be - but it's the same for every binary blob driver out there.

No: The kernel driver is opensource, so the only issue we could have is bugs in the 3D driver, the rest of the system is not affected.

If we want ARM and a GPU that has FOSS drivers, then Snapdragon is the only option. If its perf/W curves are better than, or close enough to those of the other two SoCs, I would go for Snapdragon.

Yes and no.

Qualcomm doesn't support Linux as good as TI.

The community does, but we don't know yet if everything the SoC can do is being covered by the community.

Something we need to check, that's for sure.