Windows 8 means the end of the WinNT era, says Gartner

As Microsoft prepares to finish up Windows 8 for its "Release to Manufacturing" build (reportedly sometime in mid-July), the research firm Gartner predicts in a new press release today that the launch of Windows 8 later this year is the end of Microsoft's WinNT programming model and the beginning of the WinRT (Windows Runtime) model (not to be confused with Windows RT, otherwise known as WOA).

While most Windows users will continue to run applications based on Win32 for a number of years, that will slowly change as Microsoft moves to WinRT. Steve Kleynhans, vice president for client and mobile computing for Gartner, says:

Windows 8 is more than a major upgrade to Windows — it's a technology shift. We don't see technology shifts too often; the only other one Microsoft's client OS has gone through was the move from DOS technology to Windows NT technology, which began in 1993 and took eight years, ending with the introduction of Windows XP in 2001.

Gartner believes that by the year 2020, enterprise users of Windows, along with consumers, will spend less than 10 percent of their time on Win32 desktop apps. By that time, most users will be working on WinRT-based apps running on Microsoft's Metro interface. Gartner adds, "Eventually, most Win32 desktop applications are likely to be run using server-based computing (SBC) or from hosted virtual desktops."

Source: Gartner

Report a problem with article
Previous Story

Google and LEGO team up for virtual LEGO project

Next Story

Interview: We chat with Goldhawk Interactive about Xenonauts

101 Comments

Commenting is disabled on this article.

Holy Idiot?

Does this officially break world record on how many ways could something be misunderstood in one article?

NT <> Win32 Subsystem
NT <> Win32 Kernel
NT <> Win32 API Framework
NT <> WinNT
NT <> Programming model
Win8 NT <> Architectural Change
Win8 NT <> Technology Shift
Win8 NT <> OS Model Change
Win8 NT <> Kernel Change

Windows 8 NT is not significantly different than Windows 7, anymore than each version was difference from the previous version of Windows NT.

It is still NT, it is still the same OS model, it is still the same code base.

The difference? None really, but with ARM support, people get into a kick that have no technical understanding, and do not realize that the ARM version is just a modified HAL for ARM and a recompiled version of the NT source code. (There are a few code changes in NT itself for ARM, but they are tiny, as all the portability work is done in the HAL that provides conversion and translation for the NT code base.)

What people do NOT understand is that NT has been available on many platforms, and even TODAY when you install the x86 version or the x64 version, they are as different from each other as they are from the ARM version. This has to do with the HAL of NT and how it deals with portability, and how it deals with the architecture and optimization.

So if someone were to say, the Windows 7 64bit edition was a new and different OS compared to Windows 7 32bit edition, people would laugh at them. Yet when they say the SAME thing about the WOA (WindowsRT) ARM version, it is accepted as normal, and is still laughably JUST AS WRONG.


WindowsNT is an OS architecture consisting of a new OS model, kernel, HAL, client/server layer OS subsystems, matrix layering.

WindowsNT is a different OS technology from anything that came before it, and anything that we have seen since, as most OSes are still based on prior technology concepts.

During design, WindowsNT broke with traditional concepts, purposely avoid Unix and other OS models, while also avoiding various traditional kernel technologies.

The HAL of NT is a simplistic concept, but in practice is impressive. It allows the NT code based to be uniform across architectures, and it allows for highly optimized performance across architectures. NT is coded to the NT HAL, not any SPECIFIC hardware/CPU architecture. So when moving to different hardware, all that mainly has to change is the writing of a new HAL, which was around 64KB in 1992 and is 256KB with Windows 7. So this tiny translation layer is the key to how NT can easily be ported to other CPUs/Architectures and also provide amazing performance on the specific CPU/Architecture.

(In contrast Linux or OS X or FreeBSD or *insert OS here* when moved to a new CPU/Architecture, they have to be recoded to adjust for the architectural differences, which when done 'completely' is a staggering amount of dependent code changes. Like NT they are coded in a portable language, but unlike NT their actual code base is full of architectural specific code that has to be modified. However, due to the complexity of aligning the code to the new architecture, when you see Linux or *insert OS here* running on a different architecture, they are not fully optimized for the architecture as it would be HARD when dealing with the hardware complexity we have today. 20 years ago, ya it was easy, today it is not.)


I urge everyone that has any technical interest or 'pretends' to have technical understanding to go find the first edition of Inside NT from 1993, and start there, as it gives a perspective from the team designing NT and how and why NT is designed like it is, and why traditional kernels and OS models like Unix were avoided.

WinRT is a change in programming, but so was WPF/.NET that was introduced in Vista that also does NOT use the Win32 API set. This is a developer perspective change that STARTED with Vista and XAML, and moving to a API set and framework bypassing the Win32 API set and also pushing forward to have a new OS subsystem on NT that will eventually bypass the Win32 Subsystem - but the last thing isn't happening in Windows 8.

This is the DUMBEST article and Neowin Recap of an article I have seen all year.

I honestly thought John had a better understanding of Windows NT, I am startled and disappointed.

Wrong, wrong, wrong!

thenetavenger said,
Holy Idiot?

Does this officially break world record on how many ways could something be misunderstood in one article?

NT <> Win32 Subsystem
NT <> Win32 Kernel
NT <> Win32 API Framework
NT <> WinNT
NT <> Programming model
Win8 NT <> Architectural Change
Win8 NT <> Technology Shift
Win8 NT <> OS Model Change
Win8 NT <> Kernel Change

Windows 8 NT is not significantly different than Windows 7, anymore than each version was difference from the previous version of Windows NT.

It is still NT, it is still the same OS model, it is still the same code base.

The difference? None really, but with ARM support, people get into a kick that have no technical understanding, and do not realize that the ARM version is just a modified HAL for ARM and a recompiled version of the NT source code. (There are a few code changes in NT itself for ARM, but they are tiny, as all the portability work is done in the HAL that provides conversion and translation for the NT code base.)

What people do NOT understand is that NT has been available on many platforms, and even TODAY when you install the x86 version or the x64 version, they are as different from each other as they are from the ARM version. This has to do with the HAL of NT and how it deals with portability, and how it deals with the architecture and optimization.

So if someone were to say, the Windows 7 64bit edition was a new and different OS compared to Windows 7 32bit edition, people would laugh at them. Yet when they say the SAME thing about the WOA (WindowsRT) ARM version, it is accepted as normal, and is still laughably JUST AS WRONG.


WindowsNT is an OS architecture consisting of a new OS model, kernel, HAL, client/server layer OS subsystems, matrix layering.

WindowsNT is a different OS technology from anything that came before it, and anything that we have seen since, as most OSes are still based on prior technology concepts.

During design, WindowsNT broke with traditional concepts, purposely avoid Unix and other OS models, while also avoiding various traditional kernel technologies.

The HAL of NT is a simplistic concept, but in practice is impressive. It allows the NT code based to be uniform across architectures, and it allows for highly optimized performance across architectures. NT is coded to the NT HAL, not any SPECIFIC hardware/CPU architecture. So when moving to different hardware, all that mainly has to change is the writing of a new HAL, which was around 64KB in 1992 and is 256KB with Windows 7. So this tiny translation layer is the key to how NT can easily be ported to other CPUs/Architectures and also provide amazing performance on the specific CPU/Architecture.

(In contrast Linux or OS X or FreeBSD or *insert OS here* when moved to a new CPU/Architecture, they have to be recoded to adjust for the architectural differences, which when done 'completely' is a staggering amount of dependent code changes. Like NT they are coded in a portable language, but unlike NT their actual code base is full of architectural specific code that has to be modified. However, due to the complexity of aligning the code to the new architecture, when you see Linux or *insert OS here* running on a different architecture, they are not fully optimized for the architecture as it would be HARD when dealing with the hardware complexity we have today. 20 years ago, ya it was easy, today it is not.)


I urge everyone that has any technical interest or 'pretends' to have technical understanding to go find the first edition of Inside NT from 1993, and start there, as it gives a perspective from the team designing NT and how and why NT is designed like it is, and why traditional kernels and OS models like Unix were avoided.

WinRT is a change in programming, but so was WPF/.NET that was introduced in Vista that also does NOT use the Win32 API set. This is a developer perspective change that STARTED with Vista and XAML, and moving to a API set and framework bypassing the Win32 API set and also pushing forward to have a new OS subsystem on NT that will eventually bypass the Win32 Subsystem - but the last thing isn't happening in Windows 8.

This is the DUMBEST article and Neowin Recap of an article I have seen all year.

I honestly thought John had a better understanding of Windows NT, I am startled and disappointed.

Wrong, wrong, wrong!

NT, at the core, is endian-neutral by necessity, as Intel and compatible CPUs are all little-endian, whereas non-Intel CPUs are typically big-endian (this was true of the RISC CPUs that ran the original Windows NT, for example). What the HAL does is abstract a lot of these differences from applications so that, by and large, porting became a minor chore - not a major one. One feature that Windows 2000 on x86 borrowed from RISC-based NT was a hardware emulation layer (or HEL) - this let games that expected a particular version of hardware (and DirectX) run on it (Unreal Tournament was, in fact, a major beneficiary; it ran better on Windows 2000 compared to Windows 98SE - on the same hardware). The HEL is still there across all the NT-based OSes today, and not alone Windows on desktops; it's why the same application can change so little between platforms that support the same API.

Metro is going to have to change a lot between now and then, if that's the case. It's great for what it has been designed for, but it's wholly unsuitable for a large range of productivity applications. Sometimes I wonder if the analysts are Gartner are technically ignorant.

What bull****, the WindowsNT kernel just got a whole lot more exposure because it got ported to ARM (and indirectly, Windows Phone).

I think what they're trying to say that this is the end of the Win32 era.

Gartner just make what is the most obvious "prediction" that a 10 year old could make about Windows looking at current trends. So why does anyone pay this agency for making "predictions"?

xpclient said,
Gartner just make what is the most obvious "prediction" that a 10 year old could make about Windows looking at current trends. So why does anyone pay this agency for making "predictions"?

Except a 10 year old might actually know the difference between an API and a kernel.

Gartner doesn't understand what WinNT is.
Yes, Windows 8 probably is the end of the NT kernel and Windows 9 will probably be Midori-based, but Win32 has nothing to do with NT. It just so happens that the current implementation of Win32 runs on top of NT.

Aethec said,
Gartner doesn't understand what WinNT is.
Yes, Windows 8 probably is the end of the NT kernel and Windows 9 will probably be Midori-based, but Win32 has nothing to do with NT. It just so happens that the current implementation of Win32 runs on top of NT.

Yes, Win32 has nothing to do with NT, it is both a OS API VM subsystem running on NT using the client/server constructs of NT, and it additionally an API framework. Win32 is often confused with NT because Win32 is what people are use to seeing when they start Windows. Also Win32 is conflated itself as there are two different Win32 concepts, the OS subsystem and the API framework that developers use and understand.

However, I disagree about Midori.

I think that instead of a new kernel and new OS model, what we will see are more concepts from Midori rolled into NT. There are concepts from Midori that are already have been implemented in Windows 7 and Windows 8. Also considering NT's flexibility, it could take on even more which is already happening in Windows 8 and can go further in Windows 9.

I see this move to where NT is more Midori like happening in two ways. 1) Part of NT can be moved towards managed code and being a pseudo JIT/CLR. 2) Continue to use new ways to wrap unmanaged code with realtime monitoring, virtualization, and application isolation - and other advances we have had in the past few years - that Windows 7 and 8 are already doing in various degrees.

If you take the PCA in Windows 7 today it has one dedicated feature that literally watches calls and will attempt to 'real-time' fix a bad call, and even if it can't catch it, adds this to the PCA database so that the next time the software makes the bad call Windows literally fixes it and the software runs without incident. This is one reason the stability of Windows 7 is beyond what we have seen in OSes before, because even a poorly coded piece of software running native code will get fixed on the fly by Windows or the next few times the software is run will stop failing or crashing as Windows has learned to fix the things it is doing wrong.

*PS Anyone running buggy, software, run it a few times, and see if it just doesn't start working properly/better.


So combining this realtime monitoring with sandboxing and you get the equivalent of managed code and you effectively get the end result of what Midori was doing.

This technology is already playing a role in Windows 7, and Windows 8 is bringing in new sandboxing/application isolation.

There are also other direct kernel changes in Window7 and Windows 8 that come from Midori with regard to page isolation that is reduces locks and creates walls of isolation that increase stability.


The good thing about all this is that there will always be the traditional Linux desktop when one wants to do real tech work. Although some Linux developers also try to move away from the traditional user desktop, there will always be others that will continue the development of the traditional desktop by forking projects.

Fritzly said,

I really do not think so............

I somewhat agree, though with the shift towards mobile computing with Microsoft Surface and Smartphones I do think the days of the desktop as we know it are numbered.

Luka Radunovic said,
Bottom line, Desktop is gone in next few Windows releases.

In some ways, yep. However there will be aspects of the Desktop UI concepts that will get some renewal as technology advances.

With hardware today, so much intelligence can be built into software that simply hasn't been being used like it should, there is little reason that we can't move to easier to use software UI concepts.

It is like this... What was easier, starting Word, using File - Open and digging for a document? OR Opening the folder where the document exists and double clicking it? OR Hitting WindowsKey and typing a portion of the name of the document and pressing enter?

These are two generational progressions in letting the computer and OS be smarter.

In the first jump, we no longer had to remember what software we used to create documents, this was handled by the OS, and could just open the application from the document. The next jump, we no longer had to remember where or even what we named the document.

This is the next jump, even though people don't see it yet.

neo158 said,

I somewhat agree, though with the shift towards mobile computing with Microsoft Surface and Smartphones I do think the days of the desktop as we know it are numbered.


I agree that the typical desktop user is changing: before everybody had a desktop and, professionals with specific mobility needs, had a laptop too. Now the paradigm has shifted: the average user who interact with a PC for leisure, checking emails etc. will use more and more a mobile device but gamers and professionals, people working with very large spreadsheets for example, will still have the need for large screens to look over and interacting with.

thenetavenger said,

In the first jump, we no longer had to remember what software we used to create documents, this was handled by the OS, and could just open the application from the document. The next jump, we no longer had to remember where or even what we named the document.

This is the next jump, even though people don't see it yet.

Well. it is not that people cannot see it....... the issue is to be able to deliver the concept.
What you are describing was/is the concept behind "Cairo" in the early '90 and also its next iteration: Longhorn.
Are we there yet? No we are not but... we will get there. For now I would be happy, for example, to store my emails in any folder I want and not be constrained by a monolithic ost/pst file. Vista virtual folders idea, before MS ditched them, was an interesting work-around, not a solution but still....

Caveman-ugh said,
What is the differnce between Windows 64 bit and windows rt ?

You mean Windows 8 and Windows RT... Windows RT is simply Windows for ARM processors. x86 and x86-64 software is incompatible with ARM, so your old desktop software is incompatible with Windows RT, but will work fine on Windows 8.

rfirth said,

You mean Windows 8 and Windows RT... Windows RT is simply Windows for ARM processors. x86 and x86-64 software is incompatible with ARM, so your old desktop software is incompatible with Windows RT, but will work fine on Windows 8.

You are also confusing this a bit....

Windows 8 32bit (x86)
Windows 8 64bit (x64)
Windows 8 RT (ARM)

These are essentially all the same code and OS; however, each is ported to a different architecture. The Windows 8 RT version has a few security limitations and push developers to implement a Metro/WinRT interface for software, which is the only real difference.

People would call someone crazy if they said Windows 8 32bit was a different OS than Windows 8 64bit. However, technically there is as much difference as there is between Windows 8 RT (ARM).

With the same software run? Yes and no, just like between x86 and x64 editions of Windows. x86 cannot run 64bit code for example, and if the platform shift had happened in reverse, people would have faced the same differences as they will face with Windows RT. Just like x64 code that can also be recompiled for x86, both x64 and x86 code can be recompiled to run on ARM.

(There is the artificial limitation of a Metro UI for Windows RT, that Microsoft is now requiring/asking developers to adhere because of the touch centric nature of the devices. *It may change and there may be exceptions* However, it really doesn't stop anything, as a developer like Adobe could take and recompile Illustrator for ARM, and change the UI over to Metro/WinRT and it would run just fine on Windows RT.)

TRC said,
All I know is Windows 7 is the last version of Windows I'll ever use. It was a good run though.

I am confident projects such as MATE and the Trinity Desktop Environment (based on KDE 3.5) will succeed. To me, it looks like they are the future of desktop computing.

Edited by rwx, Jun 27 2012, 4:34am :

rwx said,
I am confident projects such as MATE and the Trinity Desktop Environment (based on KDE 3.5) will succeed. To me, it looks like they are the future of desktop computing.

The future is relying on desktop environments that are forks of old versions? Brilliant. Doesn't sound like the Linux desktop is in any better shape. Old Gnome or old KDE..

Max Norris said,

The future is relying on desktop environments that are forks of old versions? Brilliant. Doesn't sound like the Linux desktop is in any better shape. Old Gnome or old KDE..

I understand the reason for your opinion but I believe you are missing the point but I don't want to start another flame war so I will just say no more as all have different opinions

rwx said,
I understand the reason for your opinion but I believe you are missing the point but I don't want to start another flame war so I will just say no more as all have different opinions

Wasn't intended as a flame post, I use a BSD desktop myself (KDE 4 or AwesomeWM) when I'm not using Windows, just curious as to how a fork of an outdated version of a desktop would be considered the future when their sole purpose is to hang on to the past. Suggesting to a Windows user that's wants to cling to an old version to "upgrade" to a different OS and then install another old desktop is just a little silly.. not only will they lose a majority of their old software but now they'll have to deal with losing new features that the later versions of these desktops bring too, never mind probably incompatibility issues with some new software. Like I said these desktops aren't in much better shape.. a fair percentage of uses don't like these new builds bringing about these forks, so now the Linux desktop starts to stagnate. Not that I can blame them.. I kind of like KDE 4 (when it works) but Gnome has gone off onto a "WTF" tangent, and Gnome Classic is a poor imitation, XFCE still feels like it was designed in 1995, don't get me started on Unity, etc.

TRC said,
All I know is Windows 7 is the last version of Windows I'll ever use. It was a good run though.

This is something I would wager against. Even if you make a dedicated choice to never use a newer version of Windows or avoid it all together, you will be using devices that will be Windows based even if you don't realize it.

This reminds me of a friend that is a major anti-Microsoft zealot, and proclaimed to me one day that his home was 100% Microsoft free. It took me less than a minute to identify his modem was running WindowsCE. There was even a device in his entertainment center running WindowsNT Embedded.

Additionally, I think you will give up the fight eventually if Windows 8 and Windows 9 pull off the success I think will happen. There are good things, even if people are not talking about them and don't see them.

Just in performance alone Windows 8 is hitting about 15% faster than Windows 7 overall, and in gaming this is even a higher percentage.

Windows 8 will be consumer release with corporate upgrade cycles hitting the Windows 9 timeframe. Which is why I include Windows 9 in my 'success' scale, as I would be surprised to see Windows 8 do as well as Windows 7 in corporate rollouts. However with the Metro and App focus, this could speed up the Windows 8 deployment for enterprise.

mutualcore said,
What I'm looking forward to is Windows Phone on Medfield chip.

Ain't gonna happen, MS ain't really pushing x86 for Mobile devices anymore, besides the battery life would be awful and the device more expensive than an Arm counterpart.

Since when did winrt become the kernel? It's a programming model not what runs the os itself. Thats like saying .net is the end of windows nt, or cars are the end of wheels. They former runs on the latter.

lunarworks said,
Wasn't .Net supposed to be the new dominant model?

How'd that go?


Well, Microsoft tried to build Longhorn on it and that didn't turn out very well. In hindsight, I think it had more to do with .NET not being mature enough, however. This was a long time before e.g. the Windows Presentation Foundation and vector based, dot pitch-independent UI's. I think the hardware wasn't quite there for the added resource requirements either. It would be interesting to see a .NET-based user environment today.

I do not see myself endorsing WinRT/Metro until apps can take advantage of a true multiple monitor configuration (multiple metro apps on separate screens).

They're assuming that WinRT equals Metro when in reality there are parts of the traditional desktop already been re-written using WinRT/XAML such as the new Task Manager which is written using said technology. IMHO wait till Windows 9 to see whether Microsoft expands WinRT to also allow the writing of desktop applications as well then we'll know what direction Microsoft is heading in when it comes to the traditional desktop.

WinRT - as it stands today - isn't a replacement to Win32. It has many merits but the lack of true multitasking, the restriction to the Windows Store and the inability to make full use of large displays will prevent it from dominance on the desktop. Undoubtedly the role of the desktop will continue to diminish and we will certainly see revisions made to WinRT but until it offers an experience that is superior to what is currently available with Win32 the likely outcome is that they will continue to coexist or will be merged together.

I don't think it is the end of the NT/Win32 era Yet. I think most apps will target Win32. WinRT is not mature and will take long time to be accepted. I am not saying WinRT can't successful but I think we have long way to go. Its going to be like the bios in PC. We all knew had to go but it took more then 15 years before we started changing to UEFI.

ctrl_alt_delete said,
I think Gartner was talking about NT based apps/Win32 apps.

which would still be incorrect as Win32 is at the moment going nowhere. WinRT is a layer on top of win32. I understand what they are trying to say, but it's strangely worded

ctrl_alt_delete said,
who knows what NT means by the way?

It stands for "New Technology." The phrase "built in NT technology" is delightfully redundant.

jeffgeno said,

It stands for "New Technology." The phrase "built in NT technology" is delightfully redundant.

NT and New Technology... it's probably a backronym.

mzta cody said,

from here- http://en.wikipedia.org/wiki/Windows_NT_3.1
"Originally, NT was targeted at the Intel i860 CPU, codenamed N10 (or "N-Ten")"

Those Wiki entries are horribly sourced. The NT page sources a set of slides (srsly?), but digging into the i860 page reveals the source to be a Paul Thurrott article from 2003 (with a broken link). Falling back to Google, here's the earliest mention of the N10 claim I can find:

http://www.winsupersite.com/ar...-one-the-early-years-127432

It was revealed by Mark Lucovsky, a guy brought to MS by Dave Cutler (if you don't know who this guy is, just understand that he's smarter than you, me, and everyone else reading this post right now) in the earliest days of NT development, so it's a pretty rock solid source. New Technology was, in fact, a later acronym to make up for the massive delay of the Intel chip pushing NT development in a new direction.

Joshie said,

Those Wiki entries are horribly sourced. The NT page sources a set of slides (srsly?), but digging into the i860 page reveals the source to be a Paul Thurrott article from 2003 (with a broken link). Falling back to Google, here's the earliest mention of the N10 claim I can find:

http://www.winsupersite.com/ar...-one-the-early-years-127432

It was revealed by Mark Lucovsky, a guy brought to MS by Dave Cutler (if you don't know who this guy is, just understand that he's smarter than you, me, and everyone else reading this post right now) in the earliest days of NT development, so it's a pretty rock solid source. New Technology was, in fact, a later acronym to make up for the massive delay of the Intel chip pushing NT development in a new direction.

Dave Cutler was the creator of VMS at Digital Equipment (DEC) - when he came over to Microsoft from DEC, he brought Jim Alchin (who would take over what became the Windows Group after he re-energized the NT subgroup) with him. DEC was best known for minicomputers and mainframes (most of which, in fact, ran VMS) - specifically, the PDP and VAX lines; examples are the PDP-8, microVAX (which was a minicomputer - not a PC) and miniVAX. (All this was before the DEC Alpha RISC CPU.)
Because of delays Intel was having with what would become the 486 line, and the rise of interest in RISC (Reduced Instruction Set Computing), in addition to Intel CPUs, what would become Windows NT also supported several RISC CPUs, such as the MIPS (later NEC) RISC CPU, IBM's POWER (a spinoff from the Apple G3, which IBM Microelectronics made for Apple) and, of course, the DEC Alpha. In fact, until the second series of Pentium CPUs came along, RISC was, in fact, eating Intel's lunch at the workstation end of business computing (due to performance primarily). What came back and bit RISC was both the second series of Pentium CPUs (starting with the 100 MHz Pentium) and Intel leveraging their fabs for all they were worth, causing massive pricing drops. No RISC manufacturer (in fact, not even the entire NT RISC group) could compete with Intel on performance-for-price with the Pentium-100 and Pentium-133 - amusingly, it was DEC's Alpha that came closest (so much so that the first non-DEC AlphaPowered workstations - from Polywell Computing - came in to supplement DEC's offerings in the workstation (read - NT) space). While Intel itself also made several RISC CPUs (in addition to the i860, which was best known for being used in coprocessing boards, Intel also manufactured the i960 and later i432, or iWarp) the bread and butter for Intel remained CISC.

ctrl_alt_delete said,
who knows what NT means by the way?

I always thought NT stood for Not Tested, learn something new every day

Ok, so lets think about this... Windows XP was Win NT 5.1, Vista was Win NT 6.0, Win7 is Win NT 6.1.... Win 8 Carries the build string 6.2... Nothing in the new build string indicates that it is not an NT based build of windows.... If Win8 was NOT an NT build, why would the version # be based on NT build strings? At the very least, if that were the case, the build number would be 7.0, or maybe WinRT 1.0.... The whole article doesn't make sense, at least in terms of its relation to the Win8 build string.

If this were a true abandoning of the WinNT foundation, the versioning would be different. Which is why I don't think MICROSOFT sees it the way "Gartner" does....

Reverend Spam said,
Ok, so lets think about this... Windows XP was Win NT 5.1, Vista was Win NT 6.0, Win7 is Win NT 6.1.... Win 8 Carries the build string 6.2... Nothing in the new build string indicates that it is not an NT based build of windows.... If Win8 was NOT an NT build, why would the version # be based on NT build strings? At the very least, if that were the case, the build number would be 7.0, or maybe WinRT 1.0.... The whole article doesn't make sense, at least in terms of its relation to the Win8 build string.

If this were a true abandoning of the WinNT foundation, the versioning would be different. Which is why I don't think MICROSOFT sees it the way "Gartner" does....

change the string and all of sudden you break hundred of thosand of applications

Reverend Spam said,
Ok, so lets think about this... Windows XP was Win NT 5.1, Vista was Win NT 6.0, Win7 is Win NT 6.1.... Win 8 Carries the build string 6.2... Nothing in the new build string indicates that it is not an NT based build of windows.... If Win8 was NOT an NT build, why would the version # be based on NT build strings? At the very least, if that were the case, the build number would be 7.0, or maybe WinRT 1.0.... The whole article doesn't make sense, at least in terms of its relation to the Win8 build string.

If this were a true abandoning of the WinNT foundation, the versioning would be different. Which is why I don't think MICROSOFT sees it the way "Gartner" does....


I can see why you're confused. It doesn't help that the article confuses Windows NT with Win32. Windows 8 and RT DO use the Windows NT kernel, but they're trying to say that most apps will be developed in the future using WinRT and not Win32.

"Gartner predicts that Nokia will push Windows Phone well into the mid-tier of its portfolio by the end of 2012, driving the platform to be the third largest in the worldwide ranking by 2013. Gartner has revised its forecast of Windows Phone's market share upward, solely by virtue of Microsoft's alliance with Nokia. Although this is an honorable performance it is considerably less than what Symbian had achieve in the past underlying the upward battle that Nokia has to face."

:-|

Well it's kind of obvious that most apps will be developed in WinRT by then. But what we should keep in mind is that WinRT is only in its first generation and is going to improve a lot over time.

But it seems Win32 is always going to run the back-end of WinRT, as well as running Web browsers and anti-virus software.

Meph said,
Well it's kind of obvious that most apps will be developed in WinRT by then. But what we should keep in mind is that WinRT is only in its first generation and is going to improve a lot over time.

But it seems Win32 is always going to run the back-end of WinRT, as well as running Web browsers and anti-virus software.

Some types of applications are utterly unsuitable for WinRT (the only reason OfficeRT has *any* traction is due to tablets/slates/netbooks without enough screen to run Office applications in a window - note that I didn't put notebooks in that category) - specifically, desktop-oriented Win32 applications, especially those that have a windowed mode (which is, to be honest, most Win32 applications). While some light *games* can (in fact *should*) be ported to WinRT, there will be some resistance even there; could you see PopCap Games porting any of their games from Win32 to WinRT? Android to WinRT? Yes. XNA to WinRT? Also yes; that would, in fact, be rather easily done. But I don't see PopCap writing a WinRT version of any of their current Win32-based games - not even Bejeweled Blitz or Bejeweled 3.)

PGHammer said,
Some types of applications are utterly unsuitable for WinRT (the only reason OfficeRT has *any* traction is due to tablets/slates/netbooks without enough screen to run Office applications in a window - note that I didn't put notebooks in that category) - specifically, desktop-oriented Win32 applications, especially those that have a windowed mode (which is, to be honest, most Win32 applications). While some light *games* can (in fact *should*) be ported to WinRT, there will be some resistance even there; could you see PopCap Games porting any of their games from Win32 to WinRT? Android to WinRT? Yes. XNA to WinRT? Also yes; that would, in fact, be rather easily done. But I don't see PopCap writing a WinRT version of any of their current Win32-based games - not even Bejeweled Blitz or Bejeweled 3.)

I haven't played the Bejeweled games, so I'm not sure whether they use any low-level Win32 APIs. To be honest, I expect even games like PC versions of Gears of War to be running on the WinRT APIs. They might only run on x86-32 and x86-64, but I'm sure the API is powerful enough for that to be possible.

PGHammer said,

But I don't see PopCap writing a WinRT version of any of their current Win32-based games - not even Bejeweled Blitz or Bejeweled 3.)

Most games use DirectX, not Win32... DirectX is fully supported in WinRT / Metro-style apps.

Meph said,

I haven't played the Bejeweled games, so I'm not sure whether they use any low-level Win32 APIs. To be honest, I expect even games like PC versions of Gears of War to be running on the WinRT APIs. They might only run on x86-32 and x86-64, but I'm sure the API is powerful enough for that to be possible.

My Mom and I both love the various PopCap games - her favorite is Zuma's Revenge, while mine is the Bejeweled series. However despite Mom having more display than I do (hers is 24", while mine is 23"), she prefers full-screen, while I play in a window. However, out of *all* the various PopCap games, only Bejeweled 3 and Bejeweled Blitz use any sort of hardware acceleration (and both, in fact, use Flash - not DirectX). However, the XB360 version (Bejeweled Live) is, in fact, a perfect candidate to cross over to WinRT (and Windows 8) - in fact, as anyone that has been following Windows 8 has noticed, several XBLA games have already made the trip to some extent (PinballFX2 in the Consumer Preview, and Montezuma's Revenge 3 in the Release Preview).

PGHammer said,
(and both, in fact, use Flash - not DirectX)

Sorry for being really corrective, but Flash uses DirectX for hardware acceleration.

Drossel said,
By the year 2020, Gartner will collapse under the weight of their own bullsh*t.

Not as long as tech companies keep paying them to find the company's product as their 'magic quadrant' leaders.

"[By 2020] most users will be working on WinRT-based apps running on Microsoft's Metro interface"

A lot can change in 8 years. The iPhone was introduced only 5 years ago. How many will even still be using Windows by 2020?

CJEric said,
"[By 2020] most users will be working on WinRT-based apps running on Microsoft's Metro interface"

A lot can change in 8 years. The iPhone was introduced only 5 years ago. How many will even still be using Windows by 2020?

If I am using Windows.. and they keep the route they went with 8.. I'll still be rockin' Windows 7.

firey said,

If I am using Windows.. and they keep the route they went with 8.. I'll still be rockin' Windows 7.


Sure. Why not. You'd be like the people that are still using Windows XP nowadays.

CJEric said,
"[By 2020] most users will be working on WinRT-based apps running on Microsoft's Metro interface"

A lot can change in 8 years. The iPhone was introduced only 5 years ago. How many will even still be using Windows by 2020?


The vast majority of people who are using Windows today.

Joshie said,

The vast majority of people who are using Windows today.
But they CHANGED THE START MENU... FUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU

CJEric said,

Sure. Why not. You'd be like the people that are still using Windows XP nowadays.

Well.... 29% of users is not so bad, is it?

neo158 said,
Gartner are ****ing idiots, end of the NT era, this is just the beginning!!!!

Me thinks someone didn't read the article.

Joshie said,

Me thinks someone didn't read the article.

Really? Tell me how NT is on it's way out... If anything it's the end of Win32, not of NT, get your facts straight!

MFH said,

Really? Tell me how NT is on it's way out... If anything it's the end of Win32, not of NT, get your facts straight!

This!
the article is nonsense, he's confusing the Win32 API and WinRT API with the WinNT kernel.
And the transition didn't go from DOS to NT, theres 9x between it (yes I know its DOS based still had a different kernel)

Joshie said,

Me thinks someone didn't read the article.

Me thinks someone needs to do more research before commenting!!!

WinRT is just the API set that runs on top of the NT kernel.

neo158 said,

Me thinks someone needs to do more research before commenting!!!

WinRT is just the API set that runs on top of the NT kernel.


Of course. I humbly admit defeat to your three exclamation marks in an argument where my only assertion was that the OP didn't read the article. You're such a better person than me. You win at life, game over, roll credits.

Or, like I said, you could read the article and realize that nowhere does the Gartner report (which, btw, we only get a glimpse of, because the full thing requires buying the right to read it) say Windows 8 is the end of Windows NT or the Windows NT era. It says, in no vague terms, that it's the beginning of the end, marking a transition away from Win32 that will take many, many years.

That assertion is very reasonable.

Equating Windows NT technology with Win32 is not ignorant. In fact, it's pretty darn historically sound. Win32 was introduced with Windows NT, and it did, in fact, take years of trudging through the hybrid Windows 9x family to get to where XP brought the entire Windows line around a common kernel for Win32 application delivery. If WinRT is the eventual successor to Win32, then yes, yes it does mark the *beginning* of the end of what defined the Windows NT era.

Gartner's full report, if anyone cares to pay for it, may have been far less ambiguous about this leap in meaning, but I can't say for sure, and I have my doubts that you guys and your vitriolic ravings are any more informed on the write-up than I am. The difference is, I'm open to the possibility that they weren't totally stupid, while you guys seemed to have come to the conclusion before you finished reading the headline.

So enjoy your knee-jerk thought processes. I'm sure it's shaving years off your lives.

Jose_49 said,

Unfortunately It's closer than we think

Like we're getting past 2012... seriously it's like these guys haven't watched the movie.

"Eventually, most Win32 desktop applications are likely to be run using server-based computing (SBC) or from hosted virtual desktops."

I really don't like the sounds of that, it sounds more like subscription based computing to me.

nate.wages said,
I really don't like the sounds of that, it sounds more like subscription based computing to me.

Yes.
When XP launched, Microsoft started showing codename 'blackcomb' prototypes and the conversation was very much subscription based computing.
With all of the changes between then and now, we see the mission hasn't changed.

deadonthefloor said,

Yes.
When XP launched, Microsoft started showing codename 'blackcomb' prototypes and the conversation was very much subscription based computing.
With all of the changes between then and now, we see the mission hasn't changed.

I've never understood opposition to software on a subscription model, as long as the cost is reasonable.

Joshie said,

I've never understood opposition to software on a subscription model, as long as the cost is reasonable.

Currently, if I run a little short on money for a few months I can continue to watch movies, write code, or do just about anything I like with my computer. If everything moves to subscription model, that will no longer be true. Mind you, they're talking about subscription based OS, turning your local computer into nothing more than a dumb terminal.

deadonthefloor said,

Yes.
When XP launched, Microsoft started showing codename 'blackcomb' prototypes and the conversation was very much subscription based computing.
With all of the changes between then and now, we see the mission hasn't changed.

Yes, I remember the roadmaps. I just hoped they had maybe changed course. I see they haven't.

nate.wages said,

Currently, if I run a little short on money for a few months I can continue to watch movies, write code, or do just about anything I like with my computer. If everything moves to subscription model, that will no longer be true. Mind you, they're talking about subscription based OS, turning your local computer into nothing more than a dumb terminal.


Microsoft does not have a history of p?ssing in hobbyist/entry-level developers' corn flakes. They would accommodate this just like dreamspark accommodates students and VS Express applications accommodate budget coders.

nate.wages said,

Currently, if I run a little short on money for a few months I can continue to watch movies, write code, or do just about anything I like with my computer. If everything moves to subscription model, that will no longer be true. Mind you, they're talking about subscription based OS, turning your local computer into nothing more than a dumb terminal.

Virtual Desktop is not aimed at consumers. It is aimed at the enterprise and even down to the small business. VMWare/Microsoft/Citrix all have compelling offerings. It's more towards virtualizing applications and secure operating environments through streaming desktops. It has nothing to do with a Subscription based OS to consumers. It has everything to do with allowing businesses to quit having to manage each desktop and start just managing a single (or however many needed) image of a desktop that is streamed to their users.

MrHumpty said,
Virtual Desktop is not aimed at consumers. It is aimed at the enterprise and even down to the small business. VMWare/Microsoft/Citrix all have compelling offerings. It's more towards virtualizing applications and secure operating environments through streaming desktops. It has nothing to do with a Subscription based OS to consumers. It has everything to do with allowing businesses to quit having to manage each desktop and start just managing a single (or however many needed) image of a desktop that is streamed to their users.

So we go back to a Pre Win95 era when there were a lot of talks about dumb terminals and everything centralized in a mainframe? I am not so sure.......... but time will tell.

MrHumpty said,
Virtual Desktop is not aimed at consumers...

You're absolutely right, I suppose it helps to read the whole article and not just the blurbs here on neowin.

It's the end of the win32 era, not Windows NT. This is the beginning of a new Windows NT boom and the birth of WinRT.

rfirth said,
It's the end of the win32 era, not Windows NT. This is the beginning of a new Windows NT boom and the birth of WinRT.

well not entirely will still have virtual win32

rfirth said,
It's the end of the win32 era, not Windows NT. This is the beginning of a new Windows NT boom and the birth of WinRT.

Yeah, WinNT is not only win32, WinRT is running on WinNT, its not a new kernel. just an updated one.
People really seem to confuse Kernel with the rest of an OS.

Shadowzz said,

Yeah, WinNT is not only win32, WinRT is running on WinNT, its not a new kernel. just an updated one.
People really seem to confuse Kernel with the rest of an OS.

In a way it's a return to the core principles of the NT kernel that MS wandered away from. It is my understanding that the original NT kernel was developed on MIPS processors even though x86 was known to be it's primary platform because Dave Cutler and crew wanted to keep them honest about multi-platform support in the kernel. Windows 2000 initial kernel development continued this trend with Alpha being it's primary development platform (even though unlike MIPS support the Alpha version was never officially released, only leaked). Since then though (starting with Win2k service packs I believe) MS has been less concerned with keeping the kernel multi-platform. Supporting ARM on Windows 8 has forced them to go back to the more multi-platform principles of the Win2k and earlier kernels. From a kernel level I think that's a GREAT thing although I'm not a huge fan of the Win8 UI on non-touch devices.

rfirth said,
It's the end of the win32 era, not Windows NT. This is the beginning of a new Windows NT boom and the birth of WinRT.

What the vice president of Gartner said was that is was the end of Win32 ("It is also the beginning of the end of Win32 applications on the desktop." were his exact words), which whoever wrote that press release turned into the end of WinNT for some odd reason.

I wonder if they'll offer a "desktop" version of WinRT without going fullscreen. The windowing system is still useful for most of everything.

KevinN206 said,
I wonder if they'll offer a "desktop" version of WinRT without going fullscreen. The windowing system is still useful for most of everything.

Agreed.

KevinN206 said,
I wonder if they'll offer a "desktop" version of WinRT without going fullscreen. The windowing system is still useful for most of everything.

I think that is the plan, but not the same kind of window/desktop system we are used to.

If you look at the concept videos they have done several times, its more about natural interfaces that are on new types of surfaces with no screens. Notice in WinRT that they remove window chrome because it feels less natural and not part of the device. Instead, I think they are going to move to an object based interface or document based so that "windows" can have different sizes, overlap, have depth, and touch all built into it. But we won't really know until Windows 9.

KevinN206 said,
I wonder if they'll offer a "desktop" version of WinRT without going fullscreen. The windowing system is still useful for most of everything.

I have to wonder who's the pharmaceutical supplier to Gartner's analysts - obviously they are getting bad lots.

The only folks that have even remotely claimed that Win32 is dead are those that are seeking to use it as a rallying cry to stop WinRT in its tracks. There are some applications that are flatly unsuitable for WinRT (windowed apps, for example); also, you CAN adopt the design standards WinRT uses without WinRT - or have you forgotten MetroTwit, which is a Win32 - not WinRT - application?

Gartner really has no clue - not even in the corporate world where many seem to value their opinion for unknown reasons.