W10M to be replaced by W10 ARM?

[Cruncher04]

My question would be why would MS put W10 on ARM if it is not the future? Does not make any sense to me if they do not plan on doing anything with it.

They will put it on Tablets/2-in-1 devices. They might put it on phablets too.
I think in mid term Windows might replace Windows Mobile on phones too. And while i see the administrative overhead a5cent is talking about, it is still beneficial for Microsoft to concentrate their efforts on a single SKU from phones to desktops. Of course for phone deployment they need to add the phone UI to Windows.

 

[Spectrum90]

Windows 10 and Windows 10 mobile are mostly the same thing. Of course, there is a segregation of features depending on the form factor in which each variant runs. Continuum breaks that segregation mixing desktop and mobile features in one device, but until now Continuum is severely limited and closer to the mobile side. For the "Creators" update Microsoft is improving the capabilities of Continuum including a more desktop like Shell. The x86 emulation on ARM adds even more possibilities.

However, the task of combining desktop features and mobile features in a converged device, like the Surface Phone, requires compromises and solving a lot of problems. I think solution could be to define profiles of operation:
1)A mobile profile that can only run UWP apps. It would provide great battery life, performance, security and low maintenance. Most people should be fine with this profile, both on PCs and Phones.
2)A mixed profile, that can run a selection of Win32 apps from the store, with security and performance constraints.
3)A desktop profile, that can run win32 apps without limitations.

The owner of the device could enable/disable the more advanced profiles when needed, both on PCs and Phones (with the hardware requirements to run Win32 on Continuum).

 

[doghouseriley]

I think it's just funny that at the official Microsoft store they refer to all windows mobile devices as "Windows Phone" devices.

 

[greedo_greedy]

Don't think it will be replaced. Windows 10 on a phone still needs a "mobile interface" after all.

It's difficult to imagine controlling Full Windows 10 on a 6" device.

My guess is it will be merged, W10M will still be utilized when on the go, but when it is docked in continuum mode then we get Full Windows 10.

The problem though is storage space. So the next mobile device will at the very least sport a 64gb internal storage.

 

[a5cent]

Don't think it will be replaced. Windows 10 on a phone still needs a "mobile interface" after all.

It's difficult to imagine controlling Full Windows 10 on a 6" device.

My guess is it will be merged, W10M will still be utilized when on the go, but when it is docked in continuum mode then we get Full Windows 10.

The problem though is storage space. So the next mobile device will at the very least sport a 64gb internal storage.

MS can easily bring the mobile interface from W10M over to W10 (if it isn't already there). You do NOT need a whole new OS just for the user interface. Bring over all the apps that are still unique to W10M, like the dialer, and you're already set.

On the 6" device you'd only ever see the same UI that we already have on W10M. Nobody is talking about interacting with the Windows desktop on such a small display. The desktop would only be available when hooked up to a larger monitor.

 

[a5cent]

My question would be why would MS put W10 on ARM if it is not the future? Does not make any sense to me if they do not plan on doing anything with it.

I guess it comes down to how we interpret the phrase "not being the future". Obviously MS would not invest into a technology that has no purpose. I'm definitely NOT saying that W10oA (Windows 10 on ARM) is pointless. However, W10oA is not the foundation on which MS is building their future. W10oA is just a means to and end. It's not the end in and of itself.

The foundation on which MS is building all their modern technology and products (Holographic OS, all their apps, etc.) is the UWP! MS couldn't be communicating any more clearly what the technical foundation for future Windows development is. However, the UWP just doesn't need W10oA! UWP software can already run on both x86 and ARM CPUs, and it can do so without requiring any additional emulation infrastructure, without the desktop, and without Win32.

As far as I can tell, W10oA solves at most three problems:

• allows "full Windows" to run on ARM chips that are far cheaper than their Intel x86 counterparts, allowing OEMs to bring Windows desktop computing to lower price points.
• allows "full Windows" to run on ARM chips that are part of a SoC which may be more suitable to mobile applicaitons than what Intel can offer (integrated GPS, motion sensors, etc)
• maintain the ability to run x86 software despite using an incompatible ARM CPU.

If it were not for the first two points, I'd maintain that Intel's Core M remains the better proposition. However, making Windows computers cheaper is always a good idea. But beyond the cost factor, W10oA is just about running legacy software from the Win32 area.

That's what W10oA does, but its goal is to solve MS' mobile OS related chicken-and-egg problem. Without apps users won't come; without users developers won't create apps. Since not enough developers will create for MS' new UWP ecosystem, MS is now providing a way for very portable devices to access the existing Win32 desktop ecosystem. MS will introduce very portable (and likely also very cheap) devices that can, when hooked up to a monitor and keyboard/mouse, run Windows desktop software. A pocketable desktop. Smaller than ever before. More affordable than ever before. There may be other surprises, but with that we now have at least one unique feature that is worth buying into despite the app gap on mobile. It could even be argued that now Android and iOS devices have a continuum related app gap.

If that catches on, and many such devices are sold, then MS will have also created a market for UWP apps, because whenever those devices aren't hooked up to a monitor and keyboard/mouse, UWP apps is what people will be using. That's what I mean when I say W10oA is just a means to an end. Its purpose is to foster a MS Windows based mobile market for which it is worthwhile to develop UWP apps. That is the real goal. Since 2013 I've been saying that Microsoft's Mobile OS (then WP8 and now W10M) needs at least one unique, highly desirable and easily marketable feature that would drive demand for MS' mobile devices despite the app gap! MS ignored that approach. Instead they tried to fix the app gap directly and (IMHO very unsurprisingly) failed. I'm glad that is now finally changing.

In the mean time, I wouldn't be surprised if we aren't given a single new piece of hardware running W10M. The casual gadget consumer may then indeed get the impression that W10oA has replaced W10M, but that impression would be wrong.

Windows has been able to run on ARM CPUs ever since W8 (Windows RT). There is nothing new about that. What is new about W10oA is its ability to also run x86 software via emulation. If MS is successful, that emulation capability will become irrelevant to all consumers except those who can't let go of their decades old software packages that have long been superseded by UWP software. That's obviously not the future. At that point, W10M also becomes a very attractive proposition (less administration and far less security vulnerabilities, smaller, cheaper, etc).

Hope that was more clear.

 

[BackToTheFuture]

Well said, @a5cent. But re.

Since 2013 I've been saying that Microsoft's Mobile OS (then WP8 and now W10M) needs at least one unique, highly desirable and easily marketable feature that would drive demand for MS' mobile devices DESPITE the app gap! MS ignored that approach. Instead they tried to fix the app gap directly and (IMHO very unsurprisingly) failed.

I wouldn't say they ignored that approach. Rather, they have always tried to leverage their desktop computing platform on mobile devices since long ago, when Windows Mobile and other WinCE-based versions started. But the hardware at the time was far from capable to run complex software, hence the birth of WinCE, which is very limited by design. Starting around 2011, ARM performance has improved dramatically. The time is ripe to bring the big fat WinNT brother to ARM, the same way they brought it to PowerPC/SPARC/Itanium platforms before. Nothing surprising really. Windows RT in 2012 was the demonstration, but it was not ready to run x86 application, and they didn't intend to as ARM was too weak, even when compared to contemporary Atom.

Along the way, they desire a flexible cross-platform software architecture that scales well to different devices of different form factors, hence the birth of UWP, which btw, started with .Net from 2000, then WinRT with Windows 8 and is streamlined in Windows 10. It takes 20yrs to make the plan come to fruition. A lot of people here think they can create software in a couple of months. For simple apps, hell yes. For foundation software like Windows, it takes years of development, trial, and redesign, then redevelopment. Windows Phone 8 was one such trial. Probably MS knew at the time WP8 was not the final plan, and did not push it hard enough to consumers. Even now W10M/W10 are not fully mature yet, but they are closely reaching the convergence point of computing platforms.

 

[a5cent]

I wouldn't say they ignored that approach.

I agree with all your points. I'm just not sure how they relate to the part you disagree with.

Ever since WP7, MS focused almost exclusively on developer/app gap related issues:

• MS payed developers to build WP apps.
• MS built WP apps for other companies.
• MS continually preached to developers that they would be rewarded if they invested into the WP ecosystem.
• Most of all, MS put a lot of effort into making it relatively easy/cheap to develop for WP/WM. More so than any competing platform.
• MS also made it relatively easy to port apps to W10M from iOS and Android.

None of that is directly relevant to users however. MS thought if they managed to fix the developer/app gap related issues, then users would follow. MS never attempted to go directly after users by giving them a reason to buy into the WP/WM ecosystem despite the app gap. That is the approach MS ignored.

There are many ways MS could have tackled that. The ability to run x86 desktop software on ARM devices is a very good way, but by no means the only unique functionality they could have provided, particularly at the time they were still trying to pitch WP towards consumers. Either way, it's the first time MS is taking this approach to tackling their mobile market share problem. Finally...

 

[BackToTheFuture]

a5cent, I agree with your opinion. But I look at it from another view point, that it was the plan to converge the platforms, THE unique feature, from long ago and it only realizes recently, and WP7/8 were not the desired outcomes. Microsoft has never abandoned mobile, they wanted it to become another general computing platform. Hence multiple versions of WM (dragging for too long), then WP7, 8, 10. WP7, 8 are intermediate products of the plan along the way, not the final pieces they were planning for. Of course they have to promote their platforms somehow, they can't leave the mobile market. They even went so far as buying Nokia, attempting a firm foothold in the mobile market. But these stopgap solutions didn't turn out well, they appeared not at the right moment when people were so fixated on droid or ios, features were missing as the OS was being developed, and MS didn't push them really hard, because probably they knew that these versions were not the final solution, and the grand design continued to change. W10 is coming close to that final solution. My opinion.

 

[a5cent]

a5cent, I agree with your opinion. But I look at it from another view point, that it was the plan to converge the platforms, THE unique feature, from long ago and it only realizes recently, and WP7/8 were not the desired outcomes.

Convergence has certainly been a general theme for MS and Windows for a long time. At least since Vista.

From 10'000 feet (sweating the details), I agree with what you are saying. However, I don't think that having had their eye on THE unique feature, suitable for stoking consumer's interests directly and estimated to release in 2016, excuses MS from NOT providing any such feature in 2012, 2013, 2014 and 2015. Like I said, there would have been plenty of other ways to offer unique, highly desirable and easily marketable features via WP.

At least in this thread, my point is really only that this is the first time MS is taking this customer oriented approach to popularizing their mobile OS, rather than going after developers/apps (and that I like :smile.

 

[Grant Taylor3]

Windows 10 on ARM will be full Windows 10 . All they will need is a Start screen suitable for a phone.

All the UWP apps for People, Phone dealer and everything else are there already.



Microsoft have spent the last 2 years or more working on Windows 10.



Getting a suitable front end for use on a phone sized device will not be an issue and unlike Windows RT we will be able to run legacy x86 applications.

They should be able to have more power efficiency and cellular connectivity that comes with the ARM platform.

Think 1 device that can be you phone, tablet, laptop with full domain join and able to fit in your pocket.

Bluetooth keyboard and mouse plus headset and USB C to connect to external monitors when needed.

With a Bluetooth headset and Cortana you would not need to take it out you pocket or bag to take a phone call.

You would also not get the Hot bag issue that you can get with intel powered kit.

 

[grahamf]

Lets turn this question around to better understand the potential:

Imagine if you could connect a small touchscreen to your desktop, like a $40 5" slate. You could use it as a second screen for odd jobs like the calculator or to look up help guides without interrupting your game, and if it connected by Wifi you could bring it around the house to manipulate what your computer is up to even when you grab a bite to eat or something.

Admittedly the use cases may be limited, but the end result is that there is more reason to make the Windows OS by default support different input methods, even Hololens and Xbox. And here's the thing: The interface is pretty much the only difference between the devices. A Surface phone may not do the same things as a Nvidia 1080 gaming system, but then again neither would a $200 2-in-1. Windows 10 actual will be a heavier OS for a phone to handle, but it makes perfect sense to come up with efficiency optimizations that even a 2-IN-1 or a hololens can benefit from.

 

[vEEP pEEP]

I guess that OS can only run on the x86 architecture. W10M - ARM.
So are we going to have two OSes with emulation on each side?

Someone smarter than me - help?

I don't understand.

One is an OS. The other is a processor architecture.

 

[a5cent]

I guess that OS can only run on the x86 architecture. W10M - ARM.
So are we going to have two OSes with emulation on each side?

Someone smarter than me - help?

W10oA (Windows 10 on ARM) is full Windows. Full Windows is comprised of hundreds of individual (smaller but generally still huge) software projects. Each project contains hundreds or thousands of files. Each file contains source code (just text). The source code belonging to a single project will contain anywhere from a few 10'000 to millions of lines of source code (lines of text). This source code is not tied to a specific CPU architecture, or at least 99.99% of it isn't, because it is generally written using a higher level programing language.

Software developers use tools, called compilers, to translate that source code into something that can be executed by a computer (because computer's don't inherently understand text files). Compiling correctly written source code into something that can be executed by a CPU requires no more than pressing a button and waiting for compilation to complete, i.e. compilation is fully automated. That's how Windows' EXE and DLL files are generated. Compilation can take anywhere from a few seconds to a few hours, depending on how large the project is.

Although modern source code generally does not, compilers do target a specific CPU architecture. MS uses Visual Studio, which is a development environment that integrates multiple compilers. One compiler will compile a project's source code for ARM. Another will compile a project's source code for x86. There are other compilers too, but we'll ignore those.

So, a developer clicks a button once and out comes W10 for x86. The developer can change one setting, press that button again, and out comes W10 for ARM. The former is just W10. Due to all the hype we call the later W10oA, but it's really just the same thing compiled for a different CPU architecture. The version created for ARM includes some extra functionality that can emulate x86 on ARM. That means when the OS loads an x86 application, it will create a run-time environment (a.k.a emulator) that will fake being an x86 CPU. The application itself can't tell the difference. Emulation is only necessary on the ARM device. Emulation is pointless on the device with an x86 compatible CPU.

This is all slightly simplified, but it's the gist of it.

 

[Krystianpants]

Completely disagree. First of all "full Windows" is not even really a thing. Nobody at MS uses that term. It is used nowhere except on consumer sites like this, due to the unnecessary confusion caused by some technically challenged people claiming W10 and W10M are the same thing. Because some insist that W10 and W10M are synonyms, we now need an additional term to refer specifically to the OS that supports the Win32 desktop environment. Hence "full Windows".

Second, there is no edition of Windows called "Windows 10 on desktop". The SKU for the desktop is called "Windows 10". Period. So says MS. It's also incorrect to think of W10 as: "W10 and W7 mixed together". At least it's no more so than W7 was W7 and Vista mixed together. What you think IS Windows 10 (the OS with UWP but without Win32) is almost everything but Windows 10. There are in fact multiple versions of Windows that ship without Win32, for example W10M, W10 IoT and Holographic. Windows 10 (without any suffix) is exactly the version that ships with Win32.

It looks like you are trying to equate Windows 10 with the UWP. That just doesn't make sense. The later is an SDK and a run-time environment for use by developers. Windows 10 is a product that includes the UWP as one of its many components

Yes I see what you are saying. As it's a transitional model in order not to alienate legacy customers. Simply windows 10 is being built over time. And this includes migrating all devs to use the new API and try to get rid of win32 or relegate it to the cloud. But you don't make such a huge drastic change all at once because no one would ever get windows 10. So they call it a service that will eventually be completely revamped little by little over time so that users don't have to take in a complete change all at once and so that both MS and devs can take their time with moving over. It's definitely a combination of things that make windows 10. One of the biggest being the kernel which was revamped. The long term goal is pretty clear. So what windows 10 is today will likely not be what it is in the future. All the stuff from its predecessors is being overhauled and replaced with the new versions over time.

 

[a5cent]

Yes I see what you are saying. As it's a transitional model in order not to alienate legacy customers. Simply windows 10 is being built over time. And this includes migrating all devs to use the new API and try to get rid of win32 or relegate it to the cloud. But you don't make such a huge drastic change all at once because no one would ever get windows 10. So they call it a service that will eventually be completely revamped little by little over time so that users don't have to take in a complete change all at once and so that both MS and devs can take their time with moving over. It's definitely a combination of things that make windows 10. One of the biggest being the kernel which was revamped. The long term goal is pretty clear. So what windows 10 is today will likely not be what it is in the future. All the stuff from its predecessors is being overhauled and replaced with the new versions over time.

Sure. However, that is exactly how MS' operating system has always evolved ever since it was first conceived. It's always been "built over time". At one point MS DOS was the OS. Windows 3.1 was just a graphical UI built on top of DOS. A decade later the roles where reversed and DOS became the command line shell built on top of Win32. More recently we had WinRT that was built ontop of Win32. Now with W10, WinRT has become the UWP and it is no longer built on top of Win32, but sits along side it, both residing above Common Core. New software layers come, really old ones go, and some layers are moved up or down within the software stack, depending on what is considered to most fundamentally define Windows. That's how Windows has always evolved.

In the last decade it was Win32 that most fundamentally defined Windows. Going forward it will be UWP (assuming it will eventually get off the ground). That's why those two components are slowly trading places within the stack. W10 is special only in the sense that it encompassed a bigger architectural change than we'd normally see from a single release (Vista was also really big). On the other hand, MS has been preparing for this step for over a decade now (Vista layed the foundation for much of W10). In the long term, no version of Windows was ever the same as it was four or five releases earlier.

All this is just to say that we really should be using the terminology in a way that is inline with technical reality now, and not with how we imagine the technical reality will be sometime in the future (a very long time from now since Win32 isn't going away anytime soon). Pretending Win32 is already gone and that W10 is only about the UWP doesn't help anybody. It's detrimental to helping people understand what is really going on under the hood, because it's technically incorrect.