Windows 8 on ARM: Microsoft bets on Office 15 and IE10 Steve Ballmer wasn't kidding when he said Windows 8 could be Microsoft's "riskiest product bet." With two-and-a-half weeks to go until Microsoft delivers the Windows 8 Consumer Preview, time has run out on holding back and Microsoft is now telling Windows devs just what they're in for with the next version of its PC client. …
"...lets you develop using C, C++, C#, VB.Net and Javascript and means Microsoft's not letting native code on Windows 8..."
C and C+ is "native code". Anything using the .NET runtime or WinRT is "managed code". Metro apps can be written using either. Metro apps can only be distributed using the Marketplace.
Reporting at its finest ;-)
The more I (sporadically) read about Windows 8 the more I wonder... Its already known that MS wants to centralize things; so "one OS to rule them all" where "all" is basically both the desktop, mobile (tablets) and phone market (a phone is mobile too of course, but you get the idea).
Windows Phone is a locked down environment. They provide all the development tools for free to do whatever you want, but the moment you actually want to use your programs you'll have to cough up some cash to subscribe for their development hub.
Yet the more I read about Windows 8 the more do I start to wonder.. Could it be that MS tries to make up for lost revenue by locking down certain aspects of Windows 8 entirely so that developers will only be able to utilize those parts in their app when they get themselves some kind of developers subscription ?
I'm not hinting at 'total' lock down, more like "locked where it matters" (or where they think it will). So; they want to push Metro forward, so everyone who wants to get their apps into Metro would now need to cough up for a developer subscription. And obviously; installation is only possible through usage of the App store which gets monitored / moderated by MS.
This is speculating on my part, but I wonder if this is where some things are headed to.
Quite risky indeed. Because since Metro replaces the start menu, how are regular developers going to get their software into this mixture would this come to pass? Dump icons on the desktop app., just like we did back in the days with Windows 3.11 and NT 4 ?
Progman.exe relived ?
I will ignore windows 8 and keep using windows 7. No windows or os x should be installed before it gets at least 2 major updates and a security update unless you want to pay for joining a beta test anyway ;) Times has changed and companies are abusing the online update scheme, opportunity to rush products to market before they are ready.
Anyway, once the Microsoft shares crash because of these Apple like actions without actually being Apple, dancing monkey is fired and windows 8.5 or 9 ships, with a start menu
We, users can have evil, long term plans too.
was from someone who thinks I really use Windows in secret !!
I don't need or use WIndows.
I browse the web, watch video, process RAW photo files, edit video ( 1080/50p), design pcbs, write code, model proteins, run extensive scientific calculations, scan and print ALL without the need for WIndows
That would make sense, but if .NET is any guide then the porting process will *not* be easier than porting to one of the several cross-platform frameworks that exist.
One issue is how much code you need to port. Is it, for example, possible to port module-by-module, relying on Microsoft to provide suitable glue to maintain a working product at each stage of the transition. This is something that they singularly failed to do with .NET, where the only "interop" provided was for OLE automation interfaces (as used by VB) rather than COM interfaces generally (as used by C/C++). Consequently, there was no evolutionary path for C++ developers and so none of them made the effort.
Another issue is whether your shiny new WinRT-based components will run on "legacy" platforms. If MS don't offer WinRT on Win7, the choice facing developers is "port to WinRT and maintain two versions of your codebase" or "choose a different porting target". If you've gone for the evolutionary (module by module) strategy, the continued support for Win7 (and perhaps even XP) is even more important because you still have paying customers on these platforms even if Microsoft don't.
You know what, I think it could be relatively easy to port virtually any Windows application to Linux, at least for rev 1
Simply take a fork of Wine, tailor it to ensure it works perfectly with your own product and distribute the whole thing as a preconfigured whole.
If you do the job properly the end user will never need to know that it is Windows code wrapped in a custom Wine "emulator*"
This gives you time to eventually port it properly to native Linux libraries.
* Yes I am aware that Wine Is Not an Emulator
You wouldn't believe the opportunity and cost savings, image overhaul companies miss when they don't do the scheme below:
1) port to gnustep/linux
2) cleanup and get rid of accidental Linuxisms compiling for freebsd
3) spare a week for osx port, yes once you start with gnustep, not the cocoa, it is damn easy. There are apps you can compile from single source resulting in osx and Linux binary.
Once you do these, ios and android is there, waiting for you. Once you get a mobile interface of course.
because Linux has never hit mainstream and never will in the desktop until they make something that doesnt need computer geeks like ourselves to sort them out.
Because Apple are very expencive devices and average joe cant afford one in the hope that developers then make software for it
Because Android is a Mobile platform that going out of control and in its current state would be daft for a desktop OS
All three of those have their places, all three are very good at what they do, but you can dislike Windows as much as you like, but fact remains it is very easy to use, it is cheap and there is software/hardware on the self that will work with them
I also remember the days we had to do chmod /dev/dsp 666 just to get sound working.
I use win 7 and have setup debian to manage my android easier, inside a low (960 mb) restricted memory, no access to actual hardware virtual machine. It isn't ubuntu, just plain debian which isn't really end user thing.
Man that was the day I really regretted my windows license, especially after watching how easily other users (the ones forced me) use it with performance.
Looks to me like the ARM version of W8 is aimed at people that might have bought an ipad or an android tablet, and the x86 version is aimed at people that might have bought a laptop.
So there's a good choice there - long battery life and low weight at the expense of compatibility/power, or vice versa.
Looking forward to seeing what the OEMs come up with on the ARM side of things - I fancy a lightweight tablet with a detachable keyboard please!
Another aspect of this is the WOA is probably going to be pretty much infection-free for a while as a lot of existing malware won't be able to target the x86's vulnerabilities directly, and of course the 'app store only' model for installing software will greatly reduce the opportunities for Trojans being installed. Add to the the locked boot-loader against rootkits, and from a AV vendor's point of view its looking like lean times ahead.
Which is good for consumers, even though I feel dirty at giving MS the thumbs-up on this aspect.
However, I guess if WOA takes off then it will be targeted by the crooks and then it will be interesting to see how long it takes for things like the boot-loader and software installation to be cracked (which is good for my penguin-fancying tendencies).
Finally, if I were a Windows developer, I would be mighty pissed off by MS allowing their own non-Metro software on WOA (Office 15 & IE10) and not allowing anyone else to do so. Wonder if OpenOffice/Chrome/Firefox or similar could mount an action against MS for such anti-competative rules?
"I would be mighty pissed off by MS allowing their own non-Metro software on WOA (Office 15 & IE10) and not allowing anyone else to do so."
Somehow I don't see Office 15 for ARM being a 'non-Metro' application. Same with IE10 (no plugins? Proceed at flank speed on that one! May I get that for x86 / x64?)
"if WOA takes off then it will be targeted by the crooks and then it will be interesting to see how long it takes for things like the boot-loader and software installation to be cracked."
The same crooks are chipping away at iOS and iDevices right now, and not succeeding as vastly as they'd like. The only difference is pundits praise Apple for locking down the device and will denounce Microsoft for doing the exact same thing.
As far as I'm concerned there are very few, if any boot-sector viruses that exist today. Exploits in IE10 will still be a vector to exploit. Software viruses really won't care about the hardware they run on. If it proves to be otherwise then I'm certain no AV vendor will be happy about this, but then again I seriously doubt that everyone is going to shift to ARM with UEFI secure-boot overnight. As far as anyone knows it's still making use of NT kernel, and it's still hybrid, meaning the same kernel-level vectors can be expected in both the ARM and the x86 hardware.
The anti-competitive behaviour came to mind, but then again that's to be expected.
"target the x86's vulnerabilities directly"
Ermm, x86 is *intrinsically* no more vulnerable than any other processor architecture. People can and do write pretty secure operating systems for the x86. I'll let their various fanbois name names, but all the world isn't Windows you know.
Of course, if WoA fails to acquire any significant market share then the black hats won't *bother* to write viruses for it, but that's another matter.
My comment was not to say ARM is more secure than x86, just that the several million x86 viruses for Windows that are already out there will, in the vast majority of cases, be inherently ineffective on the ARM version.
So the bad guys will have to port them to ARM as well and learn the new hardware in depth. Hence I expect win8 on x86 will still be targeted a lot (easy porting, majority of users adopting it with new PCs), but the WOA will not be troubled much until (or if?) it has a big enough share to be worth it.
I predict that Windows 8 tablets will sell big initially, then people will find that a "complete experience" must include a keyboard and mouse. Sales of aftermarket keyboards and mice will spike, and tablets that don't have this capability will become shelfware. Users will come to the realization that Windows 8 on a tablet is marginally useful for playing music or videos, but any real work will require the other parts that make the tablet into a desktop computer. And for some reason users will just accept this. The rest of us will continue to buy Android or iOS.
This is epic...
two versions of Windows 8 that are totally different and totally incompatible with each other.
Do they REALLY think consumers will be able to work this out? (or perhaps they are just hoping they find out after they open the box that says "no refunds")
I haven't seen the Office source code, but it exposes a huge amount of its innards through VBA so I doubt you *could* write a clone around WinRT. They could certainly write *an* Office suite that used a compatible subset of the new file formats. But then, Open Office has offered exactly that on ARM for some time and failed to set the world alight.
WinRT only exists because Microsoft want to create an API that lacks many of the features of the Windows API. That means that WinRT's designers are people who know all about the Windows API and have decided that certain sections should be withdrawn. If your program uses any of those sections (and Office's abuse of the API is the only reason that some parts of it are even documented, so it is surely such a program) then I'm sorry but porting is "impossible by design".
in the article: "We already knew that existing x86 apps would not run on Windows 8."
You did?
I must have missed that one then. Fair enough. Mind you I'm sure I've heard the same story before about previous MS OSes and it's turned out to be not 100% accurate.
As itzman said, and as any sensible developer would say in circumstances like this:
"If one is faced with re-porting all ones apps why would one not port them to Android, OSX and/or Linux instead?"
That should of course say.
"existing x86 apps won't run on Windows 8 tablet"..
In simple terms (and Microsoft have been avoiding putting this in simple terms, for obvious reasons, as confusion is how they are hoping to sell this)
Windows 8 on the PC will run "Legacy" Win32, .NET + WinRT/Metro apps.
Windows 8 Tablet will run Metro Apps + Office/IE.
Some tablets will use x86, and will have a desktop that runs legacy Win32.
That should work fairly well when mouse and keyboard are available (eg while docked). When undocked, how well it works will depend on how well the apps you use adapt to not having mouse and keyboard. It sounds like we can expect Office 15 will have been updated to support that, and probably other apps will take that route too. It's easier than rewriting for Win32, a sort-of half-way house.
ive been saying this for a while now, buying apps is all very well on mobile devices now, but what happens when you et fed up with (insert os name here) your screwed, held to either a model you dont like any more or have to bin all your apps and waste the money
unlike many on here im quite open to change if it suites my needs but getting locked down isnt cool an may cost you money in the furture.
Whats needed is a universal App store, but that will never happen.
Existing Win32 (and Win64) apps will run on the 'desktop' persona or whatever they are calling it of Windows 8 i86 and x64 editions, I don't think anyone informed has claimed otherwise.
My personal view is that no-one will bother with Metro, it will be quietly dropped and go the way of [name some Microsoft APIs etc. swept under the carpet].
As to whether a Win32 app will be easier to port to Metro WOA than Android (say), I guess it depends what you wrote it in. I suspect for most people it will be the case of keep the business logic and write a new UI, so why bother writing the UI for a platform with dubious prospects and zero market share, and produced by a company with a somewhat ahem checkered past with view to sticking to new APIs?
Indeed, there are just two reasons.
Number 1 is that you already have Win32 and sticking with it is zero effort. Porting any successful (middle-aged, fairly large) source code base to a completely new platform is a lot of work with zero short-term customer benefit. That's a really hard task to build a business case for.
Number 2 is that your customers already have Win32. Port to Metro and you've just lost 90% of the installed base of people to sell to. In fact, you'd be better off porting to Linux. Right now, Linux has a larger market share than Metro. Hmm ... looks like another Hard Business Case.
Not long time ago I jumped to Linux way of system programming, and it's much more logical than Win32 (for home projects I'm still using Win32). Having brief look at WinRT it seems Microsoft mercilessly gutted Win32 and left only what it wants to be considered "good" bits.
Quite excited to see how this resembles modern system programming paradigms (as embodied in modern OS, like Linux). E.g. Windows 7 already has condvars, symlinks and hardlinks, Microsoft is bringing new robust filesystem to compare against ZFS, so there is a hope Microsoft learned even more for Windows 8 !
Anyway, I'm not feeling sorry for developers who got too attached to Win32 and will whinge and cry about it no longer being available (on ARM tablets only!). You should have given some thought to portability earlier, or just don't target a new platform!
Paraphrasing: "one of the reasons why virtualisation won't be supported on WOA is because of battery life". Codswallop! Virtualising x86 would be a lovely little differentiator for companies wanting to offer it as an optional extra especially. Of course, running the whole stack through an x86 emulator will hammer batteries but AMD is openly pimping mixed cores with HSA and good systems management would really reduce the power drain for the couple of apps written for x86 that haven't been cross-compiled. The real reason for disabling emulation is a sop to Intel.
I smell either a huge opportunity for anyone who provides an environment for Windows x86 on ARM, presumably through hardware, or a mega-fail. Cross-compiling core components for x86 and ARM is very important to MacOS/IOS migration strategy.
You might assume hardware, I wouldn't necessarily think so.
DEC's FX!32 software almost twenty years ago allowed Win32/x86 apps to be translated (JIT style) to run on NT/Alpha instead of NT/x86. Drivers were a problem for NT/Alpha, but MS have ensured that Windows/ARM will come with all drivers necessary.
QEMU does the reverse, allows an [emulated] ARM processor to run on an x86 box. But it wouldn't take a genius (well, OK it might, but there are some around) to write an x86 emulator to run on ARM. In fact didn't Fabrice Bellard (Mr QEMU and Mr FFMPEG) already write an x86/PC emulator. In JavaScript? Yep, he did:
http://www.theregister.co.uk/2011/05/18/javascript_pc_emulator/
This will all be 'interesting' to watch. It might not play out the way MS want. It seems very unlikely to play out the way Intel would want.
> MS have ensured that Windows/ARM will come with all drivers necessary.
Are you asserting that drivers for [random old printer] or [random graphics pad] will be available when WOA ships ?
Or is it 'necessary' is a restriction that means that buying new peripherals will be necessary to suit what drivers are actually available ?
Fair questions.
"Are you asserting that drivers for [random old printer] or [random graphics pad] will be available when WOA ships ?"
Only an idiot would make that claim. I may be one, but I am not making that claim.
"buying new peripherals will be necessary to suit what drivers are actually available ?"
That would suit the industry (if not the punter) but I am not yet sure it applies to Windows on ARM.
Windows on ARM appears to be a "walled garden" environment, for both hardware and software.
Any hardware which is not Microsoft approved is not provided (and therefore cannot have a driver) because MS says you cannot need such hardware.
Any clearer?
No, I don't have a clue what they're playing at either.
It's like Apple, but without the money-to-burn fanbois.
im kinda interested in x86 Win8 on a tablet but wouldnt line up at a door and would wait for a few months or second gen hardware, this ARM bussiness doesnt sound helpful, and Im not convinced by Win8 on a desktop yet, but assuming Win8 x86 will end up on a tablet and i can do as i please on it id probably get one
WOA (sounds like what you yell to a horse just before it goes off the cliff) will
1) Will not run any existing applications
2) The tablets will probably cost just as much if not more then an IPad
3) Is based on an interface that claim to fame is it has driven down MS share of the mobile market share.
4) Is a totally different experience then what MS currently offerings (how you purchase applications, what you can do with it, ...) that is looks nothing like what they sell today. Why don't they just call it "Ms IPad".
5) Offers (as far as I see it) zero compelling reasons to buy over the current market owner (other then for the some percentage of Reg readers who either hate Apple or love spending their nights working on spreadsheets).
Someone at MS took the list of five top ways to lose the tablet race and turned it into the feature set for WOA.
WOA will live forever - in many a text books that are going to be written about how badly MS screwed this up - this is going to make "New Coke" look like the best marketing move in history.
The desktop on ARM is obviously an afterthought. Why is it there? Because the MS devs had to realize that their shiny new approach does not work at all for certain tasks, office applications being among these.
So they were forced into a compromise: "Okay, here you go, you can run Office on the desktop, but nothing else! When you're not using Office, please pretend that the desktop doesn't exist! Pretty please!"
Metro as it is right now is not suitable to become the only UI for Windows, not only on desktops (haha), but apparently on tablets as well. Considering this, the way the desktop is treated in Win8 will come back to haunt Microsoft VERY soon after the release of the OS, if not sooner.
I think it's there because they didn't have the resources to replace its functionality with Metro versions. They'd have had to rewrite all of the Control Panel bits, as well as the file explorer and Office, to use WinRT. That's much more work than recompiling for ARM and tweaking the few places that waste battery.
If WOA succeeds, they will eventually get around to rewriting the desktop functionally to use WinRT and then drop the desktop from it entirely. Meanwhile they've tried to limit the damage it can do by only allowing their own apps (and not all of those). If they truly believed mouse and keyboard were essential for certain tasks, they'd open the desktop to everyone, not just their own stuff; to do otherwise would cripple the platform.
I know a huge digital pay TV network invested millions of dollars for silver light Drm inside browser. Not naming them and they aren't in EU or USA.
obviously, they can get some kind of shell app but it breaks the basic reason why they went with silverlight, to sell content to facebook/ twitter addict, never writing browser types and of course, badly administered working types.
Some people even forgot how to install apps. Activex model was working for them.
anyway, they shouldn't have trusted Microsoft who didn't even release an official plugin (that monkey's clone doesn't count)
Microsoft's strategy is still directed towards the cloud (eg Azure). The tablet is a cloud access device so WOA essentially becomes the hardware access device to Microsoft services (including App in their App store). You tie down the browser to ensure the use of apps.
So Microsoft has a story for corporates. The corporate can move to tablets - the tablets can be x86 if you want existing code to run, or ARM-based if you want to move to cloud-based development. In both cases you have the Windows UI. So in one way it's partly shoe-horning in x86 tablets into the ARM space, but it's also styming full scale move to Android/iOS ("Wait, wait, Windows is coming on ARM..."). Microsoft is therefore trying to play two games - entry into a space where it currently has nothing, but also providing a bridgehead for x86 tablets. I just don't know if it has a real compelling reason to buy. Do you really want Excel on a tablet?
> also styming full scale move to Android/iOS
When Netbooks started they were cheaply built from DVD player screens, cheap CPUs, SD memory and free Linux. While the OEMs must conform to using Windows on all machines in order to 'earn' the MS discounts and 'advertising partnership incentives' these netbooks could not run Windows, especially Vista, so were exempt from the directives.
MS brought back XP and raised the spec just so that the OEMs would be forced (through threat of discount loss across _all_ PCs) to drop Linux.
So the OEMs started making ARM 'all in one' and tablets. These again could not run any form of Windows. MS saw this as loss of control of the OEMs, the whole basis of its marketing and sales.
So WOA is announced. ARM machines _must_ be locked so they can only run WOA. This isn't about whether consumers would buy ARM tablets or PCs, it is about stopping OEMs from shipping Linux or Android in any form, and retailers from carrying them. It doesn't matter if WOA is judged useless by the consumers, they would just have to buy x86 tablets running Windows 8 because no other alternate would be available.
"Desktop" WOA might be handy for VDI. The tablet lets MS try out all sorts of things with windows and ARM.
Consider HP/Calxeda's server - it would be quite handy to fit 288 VD's into a 4U rack system.
Like Apple's "no flash" rule, the lack of emulation is to ensure only apps which take account of ARM's architecture (limitations) are run, improving the customer experience.
i don't actually know anyone who does ! Everyone moved to Open Office years ago, and now uses Libre Office. Whatever, I will be one of the millions and millions who won't touch it, but I am looking forward to a Win8 mobile - bored shitless of Apple's whiney bullshit, and while droid is good it's not great. Win8 mobiles soonest the betterest
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
