For all the cloudy proclamations Microsoft's boss Steve Ballmer has been making of late, the company's bread 'n' butter products remain on the desktop. And that's a fact that Intel, whose lengthy and prosperous partnership with Redmond recently took a major knock, knows only too well. Which perhaps goes some way to explaining …
Virtualised/emulation layer so "one OS to rule them all"?
Funnilly enough, I remember having some "what if" discussions around a one-WIndows-for-all-chips idea a few years back with some M$ staffers. The conversation actually started around Slowaris and could it be made truly compatible on SPARC, CMT and x64, then wandered off into the Transitive software technologies and how Apple was using an old form ("Rosetta") to run Mac OS on two different CPU platforms. I mentioned that Intel's EFI was becoming an OS in itself and could it be expanded to be used as a virtualising or emulating layer, so that the same version of Windows Server could be run on both Itanium and Xeon, to which the M$ staffers admitted they had looked at the idea. Maybe they've also looking at an emulation/virtualisation layer to sit on ARM and x64, so one version fo the OS and all apps can be run on either. Maybe not the most performant solution, but it would make binary compatibility a non-issue.
Virtualisation is the way
Look at LLVM. You compile your C++ app with gcc-llvm or clang and it spits out LLVM bitcode. You can take that app and run it on any platform which has the corresponding runtime support. The app runs at native speeds because the bitcode is easily translated into machine code. The app doesn't care what it's running on anymore so the OS could be x86, ARM, MIPS or anything else. I fully expect this is how Apple intend to use ARM for some laptops and if MS are smart it will be how they do it too (or something similar).
Expecting devs to build different versions of their apps for different architectures won't cut it. Even fat binaries aren't necessary when things like LLVM exist.
So Long as It looks Like 'C' On Unix
I did a little research on LLVM a couple of years back. It looks like it's just fine so long as you are coding in a C-derived language for a Unix-derived OS. Which does cover a lot of ground, (particularly for wanker-ware on play toys); but it's far from a magic bullet.
Legacy. What it is.
Windows is a legacy operating system for legacy applications. It has been for quite some time now. It's a good thing that people are starting to realize that.
Good riddance to the bully who is finally starting to get what's coming to him.
Unlike, for example, Unix?
Intel is losing market share
...and ARM is making great advances in volume and stock price. Hissy fits on Intel's part are unavoidable. Expect more histrionics as things get worse.
Guy in the square hat, cuz it is academic. :)
Both for both
For a while now my prediction has been that there'll be mobile ARM-based devices and mostly mains-powered x86-based devices forever. Crucial common apps (which is now coming to mean your browser and little else) will be available for both, in Mac resource fork kinda style. Each platform/OS will be able to emulate the other of course, so you *can* run the thing you downloaded for your phone on your home-gateway, and vice versa. Or some ANDF (such as LLVM as suggested above) might win it. Or all 3.
I think the news is that the boundary between mobile and mains powered is blurring more and more, so now they're fighting over the same slot. So of course M$ is no longer favouring x86.
It says a lot when Microsoft are promising vapourware while everyone else in the market has the first product out and are working on their second version. Apple has their second version out and are already working on the next one.
Microsoft don't bring you the future, they bring you what was new 18-24 months ago.
Mobile x86 just doesn't cut it, you simply can't build a mobile device around it when ARM delivers twice the battery life.
>Microsoft don't bring you the future, they bring you what was new 18-24 months ago.
That explains why they've partnered with Nokia, they're both on the same strategy...
A bit harsh perhaps
I'm probably being a bit harsh about Microsoft as they did launch tablets in 2001. But they completely botched the idea and just couldn't get internal teams to support it.
It seems again that they are still unwilling to use a mobile OS on the tablet and let it gradually develop. Instead trying to remould and prune the full desktop onto a tablet.
Microsoft bring you the future
it's just they bring it several years too early. Or rather, they did under Gates. Ballamer, not so much.
MS had tablets out 10 years ago.
They had touch-based devices launching when Apple was launching the original iPod- and had things recognisable as 'Smartphones' (touchscreen, able to run complex apps, data connection, tethering, cut-and-paste, Office and email integration, etc) out in 2003.
They'd bought and launched WebTV back when Apple were still in beige boxes and turtlenecks.
They provided the tools that let the PC become the most advanced gaming platform in the world- when that lost popularity, they created and sold consoles that are now better selling than the much more entrenched PlayStation range
They created an OS that's going so strong after 10 years- a trick that Apple only managed by putting makeup on UNIX- that they're having to consciously _not_ include features in updates so that there's a bit of a reason to move to the new version. And they provided free updates for those 10 years- none of this 'ooh, it's named after a different Big Cat so it's totally worth my £50 or whatever it now' malarky.
They created the Multitouch smart-surface tech _that inspired Minority Report_, which (several years later) inspired the current breed of multitouch interfaces.
They did also create Windows ME, Microsoft BOB and miss the Internet ball totally- so they're definitely not infallible. But sometimes- just sometimes- they only fail because when they go to hit the nail on the head the nail hasn't even been made yet and the wood panel they're driving it into is still a tree.
So Microsoft DID bring you the future, and you didn't buy it because it was a bit heavy and wasn't shiny enough. Or you didn't know you wanted one until it was forced down your throat by the marketeers.
What's been added to smartphones since 2007 since BillG left and the iPhone was launched? Bugger all. They've gotten a few mm thinner and have multitouch (which, let's face it, is only used to zoom in and out). That's hardly the Future, is it?
So where did it all go wrong for M$ why are they lagging behind now?
So where did it all go wrong for them, why are they lagging behind now?
The best lies ...
... contains grains of truth.
Intel is partly right: ARM-based Win8 devices will never run legacy applications natively and there will, indeed, be different software stacks for different SoC's. But the latter won't matter as long as the programmer's API remains the same. Indeed, x86-based computers also use different stacks to compensate for differences in hardware: 32-bit/64-bit, number of cores, different graphics cards, sound cards, etc. I can't see this being much worse with ARM, especially if MS sets up a minimum standard (Cortex A8 or higher, 2+ cores, NEON FPU, DirectX support,...).
Legacy applications can be made to run under emulation, which will be slower than running natively, but perhaps not by a lot: Shared library routines can run natively and processor-hungry graphics and sound will run on the graphics and sound processors anyway.
In any case, it should be clear that Microsoft is moving away from native-code applications towards .NET, which removes any CPU dependency (as long as .NET is ported to the CPU). So in the long run, Intel has no a priori advantage over other CPU vendors.
One other point, it could translate as easily as
legacy apps can run under emulation, but with processors hundreds or thousands of times speedier than were available when the legacy app was written, who cares?
Frankly I don't think you need to dig too deeply into what MS said about the Intel presentation. If MS haven't themselves decided how they are going to do things (and they seem to be saying as much) for Intel to claim that they have made a decision is factually wrong. It is misleading because if people take Intel's claims as gospel and start planning for that direction, when MS goes a different route that causes disruption. None of that get into whether Intel is correct because they've already examined similar issues at the chip level or if they are just trying to force MS's hand on the direction they decide to go in order to facilitate other decision about what needs to go on the chip.
Does it make a difference?
MS is huge in the industry. Intel is huge in the industry. That fact has scarcely altered in the last 20 years or so. I don't know which statistics to rely on, but it seems clear that they easily have 90% of the entire marketplace. And I don't really think phones or tablets or anything else have made the slightest dent in that. MS could keep doing what they're doing for decades to come and still be laughing all the way to the bank. It boils down to the fact that we're creatures of habit. Most people simply say, "what does the majority do?", and run with that irrespective of the advantages the competition might offer.
MS and Intel will sort out some kind of mobile strategy eventually. Frankly, I don't think it matters how long they take. The notion that there is real competition in the marketplace is just hype to sell news articles.
Missed the boat
Microsoft is getting burnt because its nowhere in phones and nowhere in tablets, and if it's not careful when the ARM servers come out, irrespective of how good they are, it's very unlikely the Microsoft will have anything to do with them. As we move to a world of multiple devices per person (phones, tablets, GPS, security system,standalone printers,wireless linked cameras), only one of them - the PC - has any need for Windows and that's because of Office. You'll just want to pick up you're nearest device and have access to your data on the cloud somewhere, and maybe you'll go to the PC to do the creative work - but maybe not. And guess where the herd of developers will be moving to... Microsoft has to be big friends with ARM at the moment.
Intel's caught on the other side of the same gap. It really can't afford to wait 18 months or 2 years to catch up. And having a new fabrication technology is perhaps a barrier against ARM, but ARM is continuously being moved to better fabs by it's own customers. ARM's customers also have an armoury of proven tools and models. It's difficult to see why being x86 will give Intel any benefit. Microsoft would have to do a hell of a lot of work to make x86 remotely interesting on a non-PC device (eg start up time, software resource requirements), and why would you really be interested in Microsoft Office on a mobile phone? Office sells PCs not x86. And if Intel could catch up on watts, you still have price. Intel might spend a lot of money trying to catch up in a market where comparatively margins will be too thin to make a business at it.
Eee Hee Hee! You are all wrong!
So grossly, self infatuatingly, belly button complacently contemplatingly wrong.
The winner will be those who manage to bring relevant content to the masses at relevantly affordably prices.
It aint the technology (although that is, obviously, important).
It is the content and user/consumer perception of cost worthiness.
As a ps:
The new wave to be has no legacy - none at all.
Renée James ought to know that making comments such as won't run & never will run are bound to backfire on you. I recall in the 90's being told that you would never be able to run Microsoft OS's or applications on the Acorn RiscPC. They forgot to tell the people who went off and wrote the PC Emulator - they didn't know it was impossible, so they did it!
There are numerous stories similar to this concerning Acorn developers and the ARM chip in particular. There was a relatively small band of programmers who dedicated themselves to getting the most out of the ARM chips of the day. Now, if you could find and harness those guys you could really make things fly.
- Fee fie Firefox: Mozilla's lawyers probe Dell over browser install charge
- 20 Freescale staff on vanished Malaysia Airlines flight MH370
- Neil Young touts MP3 player that's no Piece of Crap
- Review Distro diaspora: Four flavours of Ubuntu unpacked
- Sysadmins and devs: Do these job descriptions make any sense?