Just what we were missing
I'm not sure how I survived without my "small, thin, and low-powered mobile devices" having a 16 GB OS. Where do I sign up?
Microsoft's next version of Windows will run on ARM systems using system-on-a-chip (SoC) architectures from NVIDIA, Qualcomm, and Texas Instruments. Microsoft announced its platform diversification for Windows 8 during the Consumer Electronics Show (CES) in Las Vegas, Nevada on Wednesday. Windows 8 is expected sometime in 2012 …
I'm not sure how I survived without my "small, thin, and low-powered mobile devices" having a 16 GB OS. Where do I sign up?
"Conscious of the fact that it's breaking 32-years of fidelity with the Intel architecture, Microsoft padded Wednesday's news, saying Windows is not completely abandoning Chipzilla clients."
Sorry ... there was me thinking that WIndows NT 3.X and 4 ran on the Alpha and MIPS chipsets. My memory must be failing me ... I didn't realise that these were Intel chip architectures.
.. software out there will still only be compiled for x86, and x86-64. It will take quite some time to get your favorite non-MS apps on the new platform, some may never be ported.
On the plus side some 2.3million malware flotsam will be left hanging :)
My answer is still that after XP, I'm switching to Linux. Because Linux runs on more platforms than the so called engineers at Microsoft have likely ever heard of, and it always will.
What's more NT 4 had a usable GUI, not this bloated slow and convoluted nightmare that MS has dumped on us starting with Vista. Even if it had security holes you could sail a cruise ship through, sideways. Several years ago I found that if I copied the executables for the User Manager and Server Manager to an XP box on the same LAN (not domain, no trust either) that I could merrily manage any of the above with NO domain credentials what-so-ever. Yes it was patched and updated to the max.
Security holes don't get any bigger than that.
Microsoft never really lifted a finger in support of Alpha, MIPS or Power (I can't speak for IA). What it did was sell those chip manufacturers the right to provide the equipment, engineers, programmers and support staff to do the work themselves.
When the manufacturers' porting was complete, the work that Microsoft had given vague assurances to complete (e.g., Office) wasn't done.
Obviously Intel gave Microsoft a big wad of "marketing cash" to pull another OS/2 on their paying clients.
This turned out to be a great investment for Intel; it helped to quickly bankrupt DEC, after which Intel ended up with Alpha's much-smarter chip designers - the designers who are responsible for everything from the Pentium Pro through i7 (except the failed Netburst (P4) architecture).
Intel would later give Dell and other screwdriver shops similarly huge quantities of "marketing cash" in an attempt to bankrupt AMD.
You forgot the PA Risc chips that were originally supported as well.
The real issue will be how well the provide the other applications like Office...
On Alpha NT boxes ( the last to remain after the other two were dropped after just a year or two) the only native version of MS Word was a debug build found via MSDN, which had a 18MB .exe and ran like a crap, it was even slower than using the x86 version under FX!32 emulation/transcoding.
<going off topic a bit>
This has been nice for ARM share holders, as the share price is now easily 5x what the company is truly worth (currently about £5.06 per share, was under £4 a month ago and £3.50 a month before that).
ARM make pennies from each chip sold, with a lot of their income coming from selling development suit licenses. When MS do this they will almost certainly just do another Dev-Studio compiler so ARM will like only get the benefit of those extra pennies from CPU sales.
Yes those pennies will add up, but as the share currently only offers an RIO (ie dividends) of 0.45% at today's prices. I find it hard to see the hype and silly prices lasting too long, as they would have to start selling at least 10x the chips, and versions of these chips are in everything from smart phones to set-top- boxes to washing machines already, that is unlikely to happen just because Windows 8 will be around.
Indeed those early versions of NT were simply tarts when it came to getting into bed with some of the boys not on the intel team. Interestingly the 'fidelity' only seemed to emerge with PnP on Win 2000 and later versions of NT. Slightly ironic, don't you think...?
Microsoft did a lot of work on those ports - Cutler himself was helping base port Windows 2000.
Read the MS blogs, there are some great stories there.
I don't think the 2012 target (always assuming MS actually hit it) is quite the stumbling block people think: Microsoft will want to see what the market looks like when the dust settles.
Offering ARM support is a major leap, but it's interesting to note that it's a leap that will be relatively easy to cope with if you're developing using MS' managed code (i.e. .NET) tools. I wonder how long this has been on the drawing board.
Just changing the processor isn't going to change the quality of the underlying OS though.
NT 3 and 4 ran on MIPS and PowerPC architectures. Windows 2000, XP, 2003, 2008 and 2008 R2 run on IA64 as well as the Intel/AMD architectures.
What's all this business about Microsoft only being tied to one processor architecture?
Here, let me give you a bit of a clue stick :
1) Go and look at how many hardware platforms Windows Server 2008 supports. Clue : it's more than x86 and x64.
2) How many platforms did NT 4 support on release? It's more than one, and less than five..
1) I believe MS have announced the server 2008 and studio 2010 are the last Itainium supported models.
2) How long did they support non-86 OS for, and how much MS software worked on it?
I know from experience that support/commitment from MS on non-86 was a joke. Maybe they will try harder this time, but I would wait and see.
More important will be the issue of how well the user enjoys the experience when most older hardware and software is not working, or not working well, on ARM based hardware. As that has been one of MS' key advantages so far.
The other has been bullying OEMs in to Windows-only, which I suspect is the key reason for the announcement, to avoid them defecting to Linux in order to compete with Apple, and thus avoiding users questioning if Windows is actually that good or needed at all.
"Conscious of the fact that it's breaking 32-years of fidelity with the Intel architecture"
Wrong. While originally designed for the Intel i860, and running on IA-32, IA-64, and x86-64, Windows NT also ran on MIPS, PowerPC, and Alpha processors. Those 3 were definitely *NOT* Intel.
They should have brought W7 to ARM, and still, as long as they focus on Windows ui, they are lost.
Well, it's definitely in the category of "too late." Only time will tell if it is also "too little." Maybe they will yet surprise us...
VirtualPC built a full-software x86 emulator and sold it for the purpose of running Windows on PowerPC Macs, were bought by Microsoft and their core technology is used in the XBox 360 (a PowerPC machine) to run original XBox (an Intel machine) software. If it's not completely PowerPC wedded, is it possible they could use the x86 emulator the same way Apple used Rosetta, to support old x86 apps on the ARM platform?
It's also a fantastic opportunity to emphasise Win32 as the past and .NET as the only way onward.
I wonder how companies like Adobe are going to take the news, being famously bad at keeping up with software transitions.
"Wintel" died in the 1990s, when Intel joined in the anti-trust action against Microsoft. That was viewed as a betrayal within Microsoft, and the relationship has never been trusting and friendly since then.
If you compare a PC to a tablet or phone, you find that problems such as: battery life on a PC sucks, they run too hot, and they take forever to sleep and resume.
People have always seemed to assume that this is a fault of both MS and Intel.
Putting Win8 on ARM will show that, no, it is all the fault of MS.
1) It takes forever to sleep and resume because MS' algorithms for entering sleep and wake are pathetic. Windows doesn't have a "just stop" equivalent - they do all this wind up and tear down. It won't be any faster on an ARM processor, and once it gets faster on ARM, your x86 machine will improve, too.
2) Battery life sucks because again, MS cannot shut the hell up and is always doing things when there is, frankly, nothing to do. Look for Win8 based ARM machines to also have no battery life.
3) PCs run hot, because again, MS cannot shut the hell up so the processor and chipset are always churning, doing essentially nothing. If you have used an ARM based device to watch a movie, you will notice that the device will be a little warm to the touch after a while, because it is BUSY. The difference is, when you are done watching the movie, the device cools down because the OS knows how to shut up. Windows doesn't, so the device will be warm to the touch always. Frankly, it wouldn't surprise me if you will start seeing Windows based ARM machines with heat spreaders on them!
Microsoft will fail miserably here, or said differently, if they do something that makes ARM based platforms decent running Windows, the x86/x64 counterparts will equally benefit. Microsoft has no clue what they are doing, but they are desperate to stay relevant and so look around and say "gee, ARM things are out there, guess we should do that!"
...folks have already run Windows, OSX and other BSD variants, and various Linux distros on the same or comparable hardware. There are differences, sure, but your characterization of it is more than a little OTT.
Back when Apple migrated from the PowerPC Processor line to x86, Apple included a Technology called Rosetta to allow PowerPC binaries to run on x86 chippery, it was based on a technology called QuickTransit developed by Transitive Corporation. SGI used the same tech for Irix Apps when they migrated from MIPS to Itanium. Transitive's Technology was purchased by IBM in 2009 and rolled in to AIX, allowing AIX to run x86 Linux workloads without them needing to be recompiled.
Having already been through one painful CPU Architecture migration, Apples method of migrating (and not alienating it's customers... this time!) was to distribute new software in "Universal" software binaries, these effectively contained binary executable code compiled for both processor architectures within a single installed package, older apps (compiled for PowerPC and not x86) could be emulated by Rosetta on the new machines, new Apps would run natively on both new and old processor architectures.
Interestingly Apple uses the "Universal Binary" method for iPhone and iPod Apps today (some iPad Apps too).
I'm interested to see whether Microsoft may license the QuickTransit technology from IBM or use something similar in order to allow ARM based Windows 8 devices to run Applications developed for earlier versions of Windows, or whether Microsoft might do away with legacy entirely. It's possible we could see a "Universal Binary" install mode in Visual studio, because if that didn't happen then Microsoft would need to make it so that Apps are only available from an online store and just serve up the right release depending on your CPU Architecture, this will be confusing for users if the branding is not handled well as if the user experience of both the ARM and x86 variants is not identical, users will need to understand why software will run on one windows 8 device but not on another.
I for one will watch with interest as the inevitable cock up ensues in 2012 (ha... more like 2014), when Windows 8 is finally released.
Unlikely but perhaps we will see the ARM move into traditional x86 dominated niches like desktops and laptops. Time will tell. If this happens, I suspect it can only be good.
Given windows is a reality I have to deal with:
I would like to ask M$ at this point in considering making win 8 as modular as possible (ie we get much more control of the bloat, if that's possible).
I would also ask M$ to ensure we have something like the win32 api left in it, some of us still do stuff in c or c++ calling win32 and directx directly and not that net stuff.
If this prompts a few manufacturers to produce some cheap ARM-based consumer PC hardware then I'm very interested.
"Microsoft's been taking flack for years"
Flack != flak
this is a good thing, think about it for more than a minute, it means that the white box ARM vendors can finally take off world wide as they are going to need to think about seriously upping their hardware spec's as fast as possible right.
i do find it interesting though that Freescale doesnt get mentioned here... But ....
with Freescale actually introducing what people want to actually buy this Year, that being an ARM QUAD core cortex A9/NEON 128bit SIMD powered mobile device with many hours of battery use AT FULL LOAD
then You will perhaps finally be able to buy these Quad ARM cortex NEON SIMD mobile and even desktop devices within the next 12 months, after knowing about their existence for far to long and that's a good thing,.
then go and put QNX RTP6 AROS or even Linux on there if you like later :D
"The i.MX 6 series is Freescale's first ARM-based multicore SoC and first Cortex-A9 model. The processor advances the i.MX family with dual-stream 1080p video playback at 60 frames per second (fps), 3D video playback at 50Mbps, desktop-quality gaming, augmented reality applications, and novel content creation capabilities, says Freescale.
The SoC is also touted for being one of the first applications processors to offer hardware support for the open source VP8 codec.
VP8 drives the related WebM (MKV) open container format, both of which are supported in the most recent Android 2.3 release...."
"the SoC is claimed to enable 1080p video (single stream) with only 350mW consumption. As a result, the i.MX 6 series can deliver up to 24 hours of HD video playback and 30-plus days of device standby time, claims the company."
"All three i.MX 6 models are clocked to 1.2GHz, and offer the ARMv7 instruction set with Neon multimedia extensions, VFPvd16 (vector floating point graphics), and Trustzone support, says Freescale. While the single-core i.MX 6Solo offers 256KB of L2 cache, the dual and quad versions are each said to offer 1MB of L2 cache."
About time too. Intel have been dragging their feet on properly low power implementations ever since they sold StrongARM/XSCALE to Marvell.
It will be interesting to see how Apple respond. Their only offering for a low power mobile computer kind of thing is the iPad. A closed, restricted, feature poor device; no USB, memory card slot, hideously expensive 3G add-on, etc..
A Win8/ARM offering could be all that the iPad is, but with lots of features, and an its-up-to-you-what-you-do-with-it attitude, and would probably be cheaper. Surely that would make Apple's controlling attitude look commercially unattractive.
So how could Apple respond? They would have to do the same thing - offer tablets with Mac OSX on it, not the crappy locked down iOS. But surely that would mean losing complete control over what users do with their hardware, wouldn't it? And that wouldn't not be a good thing for iTunes, surely? Besides, Apple have been neglecting OSX recently.
And Android would start looking a bit rubbish as well. What would a business user prefer; a Win8/ARM device fully integrated in to your existing corporate setup (Office, Win8, etc. etc), or a Linux derived device with no Linux apps (just weird Android apps) and a terrible OS software update mechanism? The former, I would imagine. I think Google have made a terrible mistake with Android; they don't control updates of devices that people already have. That makes it unattractive as a long lived computing platform because it is terribly vulnerable to security problems. This limits it to a disposable device market like mobile phones; got a problem with it, chuck it in the bin and buy the new one.
What would be even more interesting is if MS bolt WP7's user interface on top of Win8/ARM, and offer it as a phone OS. Then they could finally ditch the woefully crap WinCE...
Any software which runs on ARM will need to be specifically compiled for it, it means that all software for these devices will need to be compile with a new development tool or compiler and Microsoft can choose the methods by which software can be delivered to these devices... it would be a dream for Microsoft to start with a clean slate where they can push all Apps for these devices through an App Store like Apple has done. It would mean that Microsoft could:
a) Control the user experience a lot more
b) Take a slice of the revenue of any software purchase
c) Potentially reduce piracy of software
The key issue will be that until there are sufficient devices in the market, it's unlikely that developers will flock to the device, meaning that the majority of software on the devices to start with will be brewed by Microsoft, so these devices will be targetted first at corporates before they are targettted at home users, although you are bound to get basic features consumers can use like a web browser, media player and e-book reader app.
"...and the ability to print."
Oooo, they're really pushing the boat out. The next big Windows platform will be able to print? Yowzers.
Printing is one of the biggest issues in all this, so it is significant. A lot of hardware can run with generic drivers, but (apart from PostScript devices) printers are a major exception, and even where support appears baked into Windows that's only because the vendors have cooperated. We expect to be able to plug any printer into a Windows machine and it to "simply work", with only newer devices needing driver disks at all, but until we can persuade Epson et al to rebuild all their drivers from the last 20 years or more, it's not going to compare.
Microsoft's risk has been developers moving out of Windows to program apps and mobile phone devices. Particularly if corporate IT depts move to developing for non-PCs. They have to bring Windows and .NET to ARM and small devices. It's just surprising it's taken so long for them to see the danger.
I thought Windows had abandoned IA64, even on the server side (and who could blame them)  ?
Other than that, yes the Windows OS on a non-x86 box is nothing particularly new. What would be new would be a *succesful* Windows (meaning not just an OS but a decent selection of apps and drivers too) on a non-x86 box.
Maybe .NET-style managed code helps with the apps side of that, but how much stuff out there has really drunk the .NET Kool-Aid? How much of that is MS's own stuff (Office? Outlook? Stupido? etc)? There's a whole world of stuff that still assumes x86 explicitly or implicitly.
VirtualPC is an interesting thought, as indeed was DEC's own FX!32 translator (from both a technology and a supportability point of view), but these are workarounds and bandaids.
Not saying it should be that way, just that's the way the MS world is today.
"Windows Server 2008 R2 will be the last version of Windows Server to support the Intel Itanium architecture. SQL Server 2008 R2 and Visual Studio 2010 are also the last versions to support Itanium. "
I think you'd find that if MS managed to get a sufficiently consistent hardware platform for Windows 8 to run on, the Linux world would be quite capable of exploiting it for OS and apps (including recompiling existing printer drivers) very very very very quickly - more quickly than MS, probably. Unless of course MS and friends manage to encumber it with some proprietary non-disclosable stuff (probably disguised as DRM/DMCA support) to stop this happening, and they wouldn't do that would they?
It's only in phones and similar stuff where today's hardware diversity is nearly inevitable that Android and the like are currently successful (and MeeGo and Limo and who knows what else are, how shall we say, not quite so successful).
Get the modern equivalent of the Advanced RISC Computing platform hardware standard (which is what was underneath Windows NT for Alpha, MIPS, and PPC) and it's just a case of someone recompiling your Linux OS and whichever of its existing apps and drivers etc as may be needed.
Many people have been doing this for years, just not necessarily names you'd know unless you needed to. E.g. MontaVista runs very nicely on various ARMs (subject to the vagaries of support from your hardware vendor).
Am I the only one who perceives the first move on the part of MS to find a way to "embrace, extend..... etc" from software to hardware?
It looks to me like the bloke in the AC 12:13 post above was thinking along similar lines (even if it wasn't visible when you wrote yours):
"encumber <the new platform> with some proprietary non-disclosable stuff (probably disguised as DRM/DMCA support)" (to lock out non-MS OSes, as per Xbox, etc).
So it must be true then.
You heard it here first?
I just thought that microsoft already has tried a "Windows XP Media Center Edition" with little reay success. With ARM support the world of cheap set top boxes (with Internet and Freeview TV plus email, and the twatter like stuff) could be a big deal.
These could be sold everywhere, as they are much cheaper to produce that a modern x86 PC, and there are lots of people who still fear computers but would happily by a new video recorder thingymajig that also plays games.
And as others have said; they would love to have a dedicated AppStore system for getting more cash from these systems.
they have a long history of it... pre-announcing something to cause OEMs and customers to put off purchase decisions
Well, first off, I don't know about "OEM partners", but OEMs in general don't have to wait for Microsoft. They already have Android, as well as their choice of *full* Linux distros (including Ubuntu) that run just fine on the ARM. Including being able to print 8-) The beauty is these aren't second-class ports -- I've used Ubuntu on PowerPC, and previously used distros on PA-RISC, Alpha, and MIPS, and if I couldn't physically see the computer I could have thought I was on x86. Since almost everything is open source, it's not like you load up some oddball platform and you have 1/10th the apps either. (Anecdotes at the bottom)
Applications will be the BIG problem. Microsoft has two choices, both pretty nasty -- 1) Expect people to port apps to ARM. This will guarantee that Windows on ARM is second-class, whether Windows itself runs OK or not, there'll be virtually no apps available. 2) x86 emulation. This is sure to have a performance penalty, virtually guaranteeing Windows is the poorest-performing choice for an OS to run on the ARM.
Anecdote: When I was on vactaion, the guys where I used to work found a PC case identical to the PC I had on my desk, stuck a PowerMac motherboard into it, put Ubuntu for PowerPC onto it, installed the few missing apps, copied my home directory over, and put it in place of my PC. They even got a USB->PS2 adapter so the same Model M keyboard was plugged in. I literally couldn't tell it wasn't my old PC until I finally rebooted, and heard that Mac startup sound -- then I was like "What did you guys do!?" 8-)
Second anecdote: Besides Ubuntu for PowerPC, I've installed Linux distros on PA-RISC, Alpha, and MIPS. I can assure you, it was an identical user experience to running on an x86 dekstop. In fact, I used to enjoy running exotic machines rather than AMD/Intel, but it got to where the install was "pop in CD for that platform. choose 'install'. OK you're done", so I actually gravitated back to AMD boxes, until something else gets a clear advantage. The ARM has a clear advantage of FAR lower power draw.
Was the author not involved in the IT industry ten years ago?