Slashdot Mirror
Microsoft: No Windows 8 ARM Support For x86 Apps
jcombel writes "It turns out Microsoft's app compatibility will be limited to one architecture or another. Yes, Windows 8 will run on your ARM tablet, but your x86 Office 2003 will not. In his explanation, Steven Sinofsky reasoned, 'If we allow the world of x86 application support like that, or based on what we call desktop apps in our start yesterday, then there are real challenges in some of the value propositions for system on a chip,' he said. 'You know, will battery life be as good, for example? Well, those applications aren't written to be really great in the face of limited battery constraints, which is a value proposition of the Metro style apps.'"
...that just killed my desire for Windows 8 on a tablet. Thanks anyway, Microsoft.
Will there be support for ARM apps on the x86 platform?
Well, this just made Windows 8 tablets totally worthless for the enterprise.
HTTP/1.1 400
Waah porting is hard and stuff!
I'm sure "SlashdotMedia" will improve on all the wonders that Dice Holdings blessed us all with
So, once again we're going to end up having to replace a metric shit ton of applications because of an OS change. Man, I love this war on backward compatibility and long product life cycles. Not only is our hardware designed for the dump, but now our software is, too. The future sure is looking bright...
This was part of the initial announcement that Win 8 would be available on ARM. Of course x86 applications aren't going to work! You think everything is going to be run through a monumentally slow VM on already underpowered (compared to x86) hardware? .NET) will work on both. Not news.
Applications written for x86 will remain x86, applications written cross-platform (e.g.
First version of windows not to be exploitable by Virus at launch... at least until the recompile.
Momento Mori
No problem, I was going to use my fingers anyway.
*Bada boom tish*
Summation 2
basically for the same reasons as to why full OSX isn't put on iPads. So tell me again the point of having full Windows on a tablet when we're going to end up being limited anyway? Starting to lose interest already.
Well, duh. Microsoft never intended to make x86 software run on ARM. Microsoft wants you to write "Metro style" apps which are written in .NET and will run on Windows 8 x86, Windows 8 ARM, and Windows Phone.
Having failed for years to sell anyone a phone that looks like the Windows desktop, Microsoft will now make the Windows desktop look like their phone. It's backwards but it's Microsoft.
I took a peek at the Windows 8 developer preview yesterday. Nothing about it makes sense when you try to look at it as a PC desktop. Everything about it makes sense when you assume that Priority One is to "kill Google" and "kill Android." Keep this in mind when looking at Windows 8.
Tired of FB/Google censorship? Visit UNCENSORED!
Android is fragmented in an almost identical way with apps being written for specific chips, such as Tegra.
Furthermore, Metro apps WILL work across both x86 and ARM windows 8 builds, much like regular Android apps will (or at least, are supposed to) work on any regular Android device.
+1 IDisagreeSoHeMustBeATrollOrAnAstroturferOrAShill
Battery life aside, I imagine it would be incredibly painful to use a desktop application on a tablet. In a roundabout way, maybe this will devs will make some effort redesign desktop apps to fit the form factor.
Erm, just compile/build/test them for ARM? .NET apps should be simpler than most.
And
In other shock news, AMD64 apps don't work on x86 !!
Just get the DOSbox core with x86 emulation in there, identify the callouts to libraries that you've got implemented in ARM to improve performance, and you're probably do pretty well.
Heck, a port of DOSbox on its own will be enough to make many GoG games work. (This is how I play "Master of Magic" on my iPad.)
Barring that, well, glue the x86 emulator into WINE and there's probably some potential.
it would be just idiotic to emulate another architecture.
Yet Apple did exactly that for the 68K to PowerPC transition and for the PowerPC to x86 transition. Microsoft is still doing that for the .NET Framework (x86, PowerPC, or ARM emulating a hypothetical processor that runs IL), and all major browser makers are doing that for JavaScript. Even Nintendo did that for its Classic NES Series on Game Boy Advance (ARM emulating 6502).
Don't you mean:
"Windows Starter, Windows Home Basic, Windows Home Premium, Windows Small Business, Windows Enterprise, Windows Azure with Synch, Windows Media Edition"? (Or some such)
All THAT times 3 tablets!
My first Journal Entry ever, in 8 years! http://slashdot.org/journal/365947/aphelion-scifi-fantasy-horror-poetry-webzine
My 100MHz Win95 computer could handle email, web browsing, and word processing smoothly. Is it too much to expect a Honeycomb tablet, with its 10x faster Tegra core, to be able to do these very simple things smoothly?
Your 100 MHz PC was also handling indexed-color pixels with no alpha blending as well as the far simpler web page layouts that were common at the time. Furthermore, "smoothly" was interpreted in context of the 0.02 to 0.05 Mbps Internet connections that were common in the Windows 95 era.
'If we allow the world of x86 application support like that, or based on what we call desktop apps in our start yesterday, then there are real challenges in some of the value propositions for system on a chip,' he said.' and as you already know, we don't like challenges'.
Well that should have been obvious. What about the MS Office port to WinARM?
Everyone knows (or should) that the ARM tablets were NEVER going to run native x86 code. The question that Sinofsky still seems to be waffling on is the status of MS Office port for ARM. They showed that at CES, but it's disposition remains unclear.
But his Build comments indicate no x86 ports will be coming to WinARM, and that only Metro programs will be accepted for WinARM. So does that mean only 3rd party ports won't be accepted or does it mean that MS Office port is also dead? This remains unanswered.
If there are no desktop ports coming for WinARM and if MS Office for ARM is also dead, will WinARM even have a desktop mode?? If so why??
I get the impression MS is not being clear on WinARM Office port, and the ARM desktop mode, because they still don't know what they are going to do and want to keep options open. But even there they fail at communications, they could just be direct and say that hasn't been decided instead of being vague.
I think there is a significant chance that Win8ARM may be a pure metro device with no legacy desktop mode at all. This may actually make the most sense for the ARM version, but MS really needs to communicate better on this.
Clearly it would be sub-optimal to x86 instruction set on an ARM chipset. But will your license for an x86 version of e.g. Office entitle you to the ARM version of said product, when released? Or will MS double-dip?
"Ahh! I see you're in that indeterminate Schrodinger state where - oh, uh
From the summary, MS has not been trying to code for good battery life. Apparently every person working at MS uses a desktop so the point of having good battery life is non-existent. Now with Win 8 for Arm, they will start looking at good mobile battery life.
Arguments about coding x86 apps to run on Arm, sure those I can understand. But x86 not being coded for battery life is an argument that basically admits they've been screwing mobile users for years. Note that they said "those applications aren't written to be really great in the face of limited battery constraints" He didn't say that translating x86 to Arm would hurt battery life.
OK, without checking he likely is a marketing guy that doesn't even understand what he's saying.
Those who can, do.
Every time I see these Windows 8 stories I keep thinking ESPN 8, the Ocho.
We need a new slashdot meme around here anyway.
Enjoy,
It's just the normal noises in here.
I'm all for MS-flaming but how could anyone possibly be disappointed or shocked by this? It's like expecting 68k Macs to run PPC binaries. You might want it, just like you want a unicorn pony, but you don't say "OMG! NO UNIKORN PONIES? I THOT WE WERE GETTING THEM!!1 Y NOT? EVERYONE ELSE GETS UNIKORN PONIES! ANDROID ARMS HAVE XEON EMULATORS, MIRCO$OFT $UX!!"
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
Are they saying they won't provide some sort of x86 emulation (like Apple's PPC emulation), and software manufacturers will need to recompile stuff for ARM (reasonable enough; good for performance), or are they actually saying that there will be no Win32 API compatibility? If it's the latter, it's kind of stupid.
# cat
Damn, my RAM is full of llamas.
In other shock news, AMD64 apps don't work on x86 !!
Which is why almost no-one sells 64-bit apps for Windows yet.
This merely reiterates the problem with Windows on ARM: most people only buy Windows because they have Windows programs they want to run, but most Windows programs won't run on ARM.
Why can't a 3rd party emulator fill in the need?
Because Microsoft will likely decline to digitally sign it, just as Apple yanked a rerelease of a Commodore 64 game from its App Store because the player could reset the emulated C64 to the REPL of ROM BASIC. People would have to pay $99 per year for a developer license in order to run the emulator.
Is everyone forgetting about Office 365? I bet it will run on Win8/ARM. Could be an interesting twist for home users and enterprise users who don't need full on Office.
And rewrite the ARM version to use a Metro front end, according to this comment. And get the ARM version accepted into Microsoft's store, according to this article.
Fork Microsoft, Fork U 2!
"Flyin' in just a sweet place,
Never been known to fail..."
Well, there's spam egg sausage and spam, that's not got much spam in it.
Whereas x86 Windows 8 will have to remain compatible with the mass of existing code out there in the world, and would therefore be more open and useful. It might have larger system requirements but if you cared about Windows at all, you wouldn't choose any other architecture. And if you don't care, why bother with ARM.
The stupid part is that it should never have gotten to this. Microsoft should, at the very least produced an architecture analogous to LLVM where devs could build and compile their apps just the once and it didn't matter if it was running on ARM or x86 or x86-64 or MIPS or anything else. When the app installed the system could compile native code from the LLVM bitcode and run that. It might not help with x86 emulation but it would offer a migration path, one far more compelling than what's on the table at the moment.
I thought they were trying to force people onto their "cloud" versions anyway.
Having to work for a living is the root of all evil.
To add x86 support would cost an ARM and a LEG.
Slow down, cowboy! It has been 4 hours since you last posted. You must wait another few hours.
ARM is a 32-bit platform, so NO 64-bit applications.
How about you make the x86 Windows apps run on the ARM tablet, and let me worry about how much battery I have. Ok?
Entia non sunt multiplicanda praeter necessitatem.
youre talking about the difference between a tablet wasting clock cycles on emulation (winARM) vs a plugged in PC (Mac). not comparable.
Mac laptops have been around since the 68K days. The PowerPC PowerBooks had to emulate 68K application code, and the x86 MacBooks had to emulate PowerPC application code. But few to none of the system library calls were emulated, especially by the time of Mac OS 9.
some emulators (VM's) take the extra step of actually compiling that bytecode into native code during execution, as do modern browsers with javascript now.
The Rosetta emulator in Mac OS X 10.4 and 10.5 for Intel Macs was likewise a JIT recompiler.
i think to do proper software emulation you would need 20x the horsepower (lets trivialize that to clock speed for now) of an emulated chip.
You'd need 20x the horsepower for the application code but 1x the horsepower for the library code because it wouldn't need to be emulated. And it'd probably be down to 3x the horsepower anyway once the JIT kicked in.
Microsoft has apparently had their Google Moment. Or maybe their Internet Moment, again.
Back in 2005 or so, Google decided that search -- their bread and butter -- was moving off the desktop and onto mobile devices. That should have been a signal that PERSONAL COMPUTING was doing this, particularly given how web-centric the average Joe has become lately. But I digress. The big problem was ownership of that desktop: while they couldn't control the web, at the time, Palm, Microsoft, RIM, and Nokia controlled the handset. It was all proprietary, and default search could change instantly, kicking Google off the mobile device overnight. So they bought Android, with the intent to put Android out there free and easy, preventing any mass defection to some other search engine on the handset.
Like Microsoft's discovery of the internet after everyone else was already wading knee deep, I think they have discovered that handheld devices are going to be significant. And that doesn't mean the PC dies. But look at it this way: Apple has 5% of the PC desktop. Linux, maybe a percent or two, Windows owns the rest. If that's the only personal computing platform, Microsoft has it pretty good.
But now enter handhelds. If 25%, 50%, 75% or whatever of personal computing moves to the handheld, it's not simply Microsoft not being there. It means that all of a sudden, the desktop is in the game again. If I have Android 5.0 on my phone and my tablet, maybe I'll want it on my desktop, too. Not a problem... the Linux roots were born on the desktop, it already runs there today (well, Android 2.x or 3.x). This is already happening, a little, with PCs... the iPhone and iPad have had a small coattail effect in the USA, boosting Apple's consumer market PCs to the extent that they're actually building volume, even as they drop some of the pro market stuff (the servers are dead, and the high-end Mac PCs are looking threatened -- Apple seems to want everything to be based on common laptop hardware: that's notebooks, iMacs, and Mac Minis alike).
But you can't put Windows on a tablet. Oh, sure, you can... they've been doing it for over a decade. But you need a stylus, a hard drive, and lots of money... like small notebooks before the netbook, classic Windows tablets are overpriced and underpowered.
But to live on an iPad/Android style tablet, whether its x86 or ARM powered, you have some issues. Like, 16-32GB of storage, tops. A CPU that's no more powerful than the Atom. Smaller displays. Touch interface. None of that stuff is going to support existing Windows applications.... particularly not ARMs, which would need recompiles for any native-code application (most of them). And complete rewrites for the ARM anyway.
So it looks like Microsoft's taking a different tack. They'll keep the x86 apps on the desktop and notebook PC, pretty much. The whole "Metro" interface comes with a totally new API, same thing more or less they run on the phone. That means you build applications, or more properly I guess, apps, in HTML, CSS, Javascript, and Silverlight. Which ought to run on ARM, x86 and 64-bit x86 all alike, maybe even PCs to phones. That's actually not Microsoft's original strategy -- that's just what HP was talking about doing with WebOS... which is also based on HTML, CSS, and Javascript. Imagine that.
So MS is inflicting the touch-like interface on PC users to leverage the vast market for PCs onto the tablet, I guess. People will buy new PCs, and as Vista taught us, they will buy them with any old OS Microsoft tossed onto them, eh? MS isn't yet worried about desktop survival, so angering a few desktop users doesn't matter if that facilitates a whole scale migration of the Windows ecosystem to smaller devices. I bet they include a Zune Store or something similar for Windows apps written Metro-style... maybe even go all Apple on us and make that the only place you can buy them. These would potentially run on all your devices, from phone to desktop. None of the hardcore desktop stuff is involved -- they're really have two completely independent application environments for the desktop systems, but only the one for tablets and phones.
At least, that's where I see it all going.
-Dave Haynie
Obviously apps written in HTML5/Javascript as well as C# are portable across architectures.
Only if the C# doesn't call native DLLs. If, say, it calls zlib.dll because C does compression faster than C#, you're SOL (yes, you could find an equivalent ARM DLL, but it won't work out of the box).
Metro apps, I think, are not x86 or ARM, not 32-bit or 64-bit either. They're interpreted... Javascript, HTML, CSS, and Silverlight only. They're light weight, intended for apps, not applications. They'll fit on phones and tablets just dandy, but also run on the "looks like a tablet" version of the desktop.
If it doesn't make sense, I think it's because this is kind of a desperation move by Microsoft. They see computing walking away from being desktop centric. If they don't follow, then take these new platforms over, Microsoft ceases to be the powerhouse they have been. And it's already too late to just launch a tablet OS... who's going to buy a not-Windows WIndows OS for the tablet, when you have Android or iOS to chose from, with thousands and thousands of apps? Look at the phone so far.. lots of good reviews. I'm not a fan, but non-geeks seem to like it. And yet, Microsoft is still losing market share in the smartphone market. Maybe they'll start gaining -- still a bunch of SymbianOS and RIM customers to steal, even if they can't get the iOS or Android users.
But desktop is their strength, and I'm certain Metro is their way of moving apps from the desktop to the smaller devices, en masse. Because developers will support the desktop, some even the new programming interfaces. MS may even have rules about when you're allowed to use the old APIs, and when you must build with Metro. I don't think they're looking at Photoshop in Metro, but other stuff, certainly. You probably have to build in Metro to get into the Microsoft Appstore, which is also a way to get traction with desktop users.
-Dave Haynie
No one sells 64-bit applications? I disagree -- most of my professional applications are 64-bit. Most still have 32-bit versions. But some in the same areas (Adobe Premiere, for example.. I use Sony Vegas) are now 64-bit only.
What you're going to see is a tier of applications. Not every x86 application needs to be 64-bit. If you don't need more than 2GB RAM, there's not much point.
Similarly, there will be Metro apps. These are not compiled, they're built in HTML, CSS, Javascript, and Silverlight. They'll work on any version of Windows, but no heavy lifting.
-Dave Haynie
No one sells 64-bit applications?
Hint: 'almost no-one' doesn't mean 'no-one', it means 'almost no-one'. Note the 'almost'.
I have three 64-bit Windows programs on my laptop that I'm aware of; IE, Eyeon Fusion and APB Reloaded. Unless 64-bit Windows comes with 64-bit Notepad, everything else is 32-bit; certainly Media Composer and all the other games are.
the full name of x64 is x86_64. Office 2010 was probably recompiled for the purpose.
Microsoft is doing the right thing here. They are dropping the antiquated Win32 API and its bloat, in favor of a new universal WinRT API that targets both ARM and X86. Furthermore, it consolidates everything (Silverlight, Win32, WPF, .Net) into a cohesive API that you should be easily able to port to. If you cannot easily port to it, you're probably designing some custom business app that has and never will upgrade. Sorry but the new Windows isn't for you.
People have been asking Microsoft to the drop the bloat for some time. The security has been a nightmare because they've had to maintain the old unsecure model for the sake of compatibility. This clean break allows them to fix the permission model so that each App asks and gets only the permissions it needs during the install (like Android).
I like this change. Apps will now have to focus on doing one thing really well (and being able to connect and share with other apps) instead of being one-stop bloatware packages. Have you seen new contract API that facilitates this? It seems to follow the UNIX philosophy of doing a single task and being to pipe that output to anywhere.
there are many companies that do it - it's not almost no one. Autodesk has pretty much their entire line in 64 bit - everything from autocad, revit, maya, inventor, etc etc etc. Those are used all over the place. Adobe's CS5, including photoshop, with premier and and after effects as exclusively 64 bit. 3d studio max, cinema4d, etc etc. Lots of them. Basically all the programs that can gain from lots of memory, have done so already.
If a vendor has to recompile an application with a few minor changes, they might do it.
But if they've got to rewrite using Metro, they're more likely to just ignore ARM.
Hell, I don't even want Metro interface apps on x86! Why would I want to be constrained to touchscreen considerations on a desktop?
I do not fail; I succeed at finding out what does not work.
vaporware. Hear that gurgling noise, Intel and Microsoft? Your boat is leaking.
Way back when, we had a 64 bit DEC Alpha server running windows NT in our geek lab. Windows software compiled for 32 bit x86 did run on it under a tool called digital FX!32 - iirc it was a combination of x86 emulation and translation to native code. The alpha was by far the fastest CPU we could afford at the time, so the performance measured up pretty well against running the same 32 bit x86 software natively on an intel based PC, while the performance of natively compiled code was (at the time) enough to make us drool. Windows NT was available for number of different CPU architectures at the time, as it wasn't yet certain that intel would become so dominant. I don't know much about the state of windows these days, but windows NT had been designed with portability in mind.
The theory and practice of emulators and running code in VM's has come a long way since. I'm sure it would be even more feasible now to run x86 software on a different CPU. Why not take this guy's statement at face value? They probably did some tests, and the result probably simply wasn't worth it. You'd be running applications optimized for a high performance PC with, mouse, keyboard, a large screen emulated on a device with a small touch screen and with slow, but power efficient hardware, and a few minutes into your experience you'd conclude that this windows tablet is just slow, awkward en inefficient and just not worth it, while you'd have a much better experience with a native app designed for it.
I don't understand all the hate on this forum about a version of windows that's not even out yet. I think it's interesting that they seem to be taking a risk and trying something radically different from what has been financially so successful these past 15 years or so.
Sorry but the new Windows isn't for you.
Not necessarily; the new Windows on ARM might not be, though.
I'd go with bytecode on the install disk and a pre-compiler that spits out native code optimized for the hardware. It would run on anything, and on every version of your os as well (though you might need to re-install when you upgrade the OS).
Microsoft already told us Windows 8 would run on phones. Did people think Win32 would be there in Nokia phones? No. They will be Metro only.
Of course on iPads, they are also going to give you Metro only. How could it be any other way? More than 99% of all tablets sold were mobile-optimized. The software on them was built 100% for mobile use, which is the use case of a tablet, and as such, enables iPads to be 1/3rd the price of a Wintel tablet, 1/2 the size and weight, 2x the battery life, and 100x more popular.
If your device has an Intel chip, you will have the "Desktop" Metro app in your Metro. Only difference. In a sense, Microsoft is committing even more to mobile, they are making all of their devices mobile, all Metro, all-new apps, and Wintel Desktop is now an app you can get on the more expensive, bulkier devices that have Intel chips.
What this misunderstanding comes down to is that Microsoft can't Osborne Windows while they get Metro up and running. They just shipped Metro development tools a couple of days ago. The false impression that legacy Windows will be on ARM is beneficial to Microsoft right now because 100% of Microsoft apps are legacy Windows apps. But their Windows 8 Store does not even sell legacy Windows apps. Only Metro apps. And their ARM devices will only run Metro and Metro apps.
Their pitch that you could have the same (Metro) interface on your phone, tablet, PC, and so on is essentially true. But the phone will have a dialer app and the tablet will not, and the PC will have a Desktop app and the tablet will not. But they will all have Metro.
are you upset that windows 7 applications didnt run on windows 7 phones?
No, I'm not upset that applications for Windows 7 don't run on Windows Phone 7 devices, because "Windows Phone" comes before the version. If Windows CE were called "Windows 95 Phone Edition", you might have a point.
instead everyone is taking it as a negative that they didnt go far enough.
The negative is that Metro is the only environment on the tablet, and Metro apps need to be signed by Microsoft, just like iOS apps on iPad. There appears to be no counterpart to "Unknown sources" on Android, and Microsoft appears not to have released details on how much Metro application development might cost.