Microsoft restarted Windows Phone from scratch

The New York Times has a new and interesting article this weekend on Microsoft's development of its Windows Phone operating system. Among other things, the story claims, via unnamed sources, that Microsoft and Nokia will indeed launch the Lumia 900 smartphone at CES 2012 next week as we have previously reported.

The article touches on the fact that while Windows Phone-based products have gotten good reviews it has yet to gain a huge audience. It also goes over the origins of Windows Phone. It states that after Apple launched their iPhone with the iOS operating system in 2007, Microsoft decided that its previous mobile phone OS, Windows Mobile, would not be able to compete with Apple's new product.

During a meeting in December 2008, Microsoft's Terry Myerson asked the mobile group if anything could be used from Windows Mobile for a revamp of the OS. After the meeting was over, the consensus was that Microsoft should start from scratch and create an all new mobile phone OS. That decision, while a good one from a programming and engineering sense, also delayed Microsoft from launching Windows Phone. 

That also gave Google an opening when it launched its own smartphone OS, Android, in 2008 and which went on to become a huge success with multiple phone makers. Microsoft launched Windows Phone two years later in the fall of 2010 with the major "Mango" update in the fall of 2011.

Myerson now heads up the entire Windows Phone group and it will be his job to convince the general public that Windows Phone devices are worth getting compared to the iPhone and the numerous Android and Blackberry products. Myserson admits in the article, "Entering the market so late with this experience has created some special challenges for us. I think if we were there earlier it would be different."

Report a problem with article
Previous Story

Rumor: Diablo 3 to be released February 1?

Next Story

Motorola teases new power related product

71 Comments

Commenting is disabled on this article.

techguy77 said,
They might start from scratch again.

haha, funny.
no way, at least that's what i assume... WP7, though i hate the name... is actually competetive now in the mobile market in terms of functionality - but it doesnt have anything too special that will make users switch from popular mobile OS's now. I do hear that Apollo is supposed to be a rather impressive update/upgrade. If Microsoft wants to gain traction they will have to introduce something that will actually make users "WOW", that wow factor can get its market share - kind of like BB previously did.

auziez said,

haha, funny.
no way, at least that's what i assume... WP7, though i hate the name... is actually competetive now in the mobile market in terms of functionality - but it doesnt have anything too special that will make users switch from popular mobile OS's now. I do hear that Apollo is supposed to be a rather impressive update/upgrade. If Microsoft wants to gain traction they will have to introduce something that will actually make users "WOW", that wow factor can get its market share - kind of like BB previously did.

I agree

While the world will have you believe that they 'restarted from scratch', it's just not that easy to build an OS from scratch.

From what I've seen from phones with interop unlock, there is some legacy locked away from the average user.

Sigh, Windows Phone was not "recoded from scratch". That implies they started from nothing, which is absolutely not true. Windows Mobile and Windows Phone are both based in the Windows CE kernel. People need to remember that kernel != OS as they know it as its much more than that.

Windows Phone is a complete restart in that they ditched complete backwards compatibility with older Windows Mobile apps and became touch first. However, that's been known for a long time.

I thought this is old news o.0
I nearly thought Microsoft is going to remake Windows Phone again from scratch after Windows Phone 7....

Is this like some revising old news trend? NYT also states "By then, a consensus had emerged (at MS) that there wasn't much technology worth saving". I bet many existing Windows Mobile users beg to differ. By doing a reset, they fell in behind in completeness of features. You don't reset everything you have built for years. You work on its problem areas only preserving compatibility and taking the best features of the old.

xpclient said,
Is this like some revising old news trend? NYT also states "By then, a consensus had emerged (at MS) that there wasn't much technology worth saving". I bet many existing Windows Mobile users beg to differ. By doing a reset, they fell in behind in completeness of features. You don't reset everything you have built for years. You work on its problem areas only preserving compatibility and taking the best features of the old.

So they should have kept the old, slow and bloated version and just slapped a new UI on it instead?

xpclient said,
You don't reset everything you have built for years.
You do if it has major problems and isn't working. And this is coming from an early Xperia (WM) user

neo158 said,

So they should have kept the old, slow and bloated version and just slapped a new UI on it instead?

I think WinMo wasn't too shabby, it's just that the hardware back then didn't allow it to enter full stride. The same goes for Symbian as well. There is so much versatility in both OSes (maybe WinMo more so when Symbian started with all the unsigned/signed **** in 9.3) which iOS is now just struggling to catch up with. There are so many things I could do with my WinMo 6.5 phone and my Nokia 5800 that I can't do with my iPhone (for example, exporting contacts to SIM card).

What Microsoft discovered too late was that the interface only told half the story: you need to have a robust ecosystem for buyers to jump on. Compared to iOS and Android, Windows Phone has a long way to catch up in that aspect, which is why people are hesitant.

People who jump from iOS to Android can make the move easily because Android can do almost everything that iOS does. Android to iOS converts don't really lose out either, plus they gain a much more robust default mp3 player in the process and also get away from the version fragmentation bs. WP is more of a paradigm shift (in terms of UI) that people are afraid to get onboard to, plus cross-platform apps generally don't enjoy much priority on WP than Android/iOS (e.g. the same app may be released months later after their Android/iOS counterparts). Hence, for the average joe they will probably stick with iOS. With WP7 you can do really much less than in WM6.5, hence the more technically-inclined will probably go for Android or even MeeGo instead.

trenzterra said,
plus cross-platform apps generally don't enjoy much priority on WP than Android/iOS (e.g. the same app may be released months later after their Android/iOS counterparts).

That's definitely Microsoft's fault, as both iOS and Android allow native code - so you can reuse your code for both platforms. WP7 is too focused on .NET and requires the developers to rewrite their programs…

trenzterra said,
I think WinMo wasn't too shabby, it's just that the hardware back then didn't allow it to enter full stride. The same goes for Symbian as well. There is so much versatility in both OSes (maybe WinMo more so when Symbian started with all the unsigned/signed **** in 9.3) which iOS is now just struggling to catch up with. There are so many things I could do with my WinMo 6.5 phone and my Nokia 5800 that I can't do with my iPhone (for example, exporting contacts to SIM card).

What Microsoft discovered too late was that the interface only told half the story: you need to have a robust ecosystem for buyers to jump on. Compared to iOS and Android, Windows Phone has a long way to catch up in that aspect, which is why people are hesitant.

People who jump from iOS to Android can make the move easily because Android can do almost everything that iOS does. Android to iOS converts don't really lose out either, plus they gain a much more robust default mp3 player in the process and also get away from the version fragmentation bs. WP is more of a paradigm shift (in terms of UI) that people are afraid to get onboard to, plus cross-platform apps generally don't enjoy much priority on WP than Android/iOS (e.g. the same app may be released months later after their Android/iOS counterparts). Hence, for the average joe they will probably stick with iOS. With WP7 you can do really much less than in WM6.5, hence the more technically-inclined will probably go for Android or even MeeGo instead.

True, I had Windows Mobile 2003 on a PDA, Windows Mobile 5 on the Omnia, both of which sucked hardware and software wise, and now have a Touch 2 with Windows Mobile 6.5 as a work phone. I totally agree that the hardware was lacking, having said that the HD2 hardware did make a massive difference speed wise. I do think that the only version of Windows Mobile I would consider using now is 6.5.1, simply because on the phone I have for work is pretty fast even though the processor is only 528MHz.

As you say the ecosystem is important as well, Windows Phone will get there but the platform hasn't matured enough yet. That's what happens when you redesign the OS from scratch I suppose.

The other problem is that most of those people who are hesitant to switch have a lot of apps already for either iOS or Android. They would have to buy again or find a similar app on Windows Phone if they switched.

I hear you on cross platform apps as well, Skype still isn't out for Windows Phone despite now being owned by Microsoft. I do think that WP has a robust MP3 player in Zune as well. Sure, the amount of content might not rival the iTunes store but it's not that bad.

Edited by neo158, Jan 8 2012, 5:16pm :

trenzterra said,
I think WinMo wasn't too shabby, it's just that the hardware back then didn't allow it to enter full stride.

This is true. What was the final nail in WinMo's coffin was the Qualcomm MS7200 processor. It was a decent processor when it first came out, but manufactures continued to use it for 3 full years after it came out. Same processor for that long while everyone else kept advancing and using faster processors.

I still kind of hope they would have released the Photon which shown in the beginning of 2008 on some decent hardware.
http://www.engadget.com/2008/01/06/is-this-windows-mobile-7/

It was an evolution of Windows Mobile and would have retained backwards compatibility and all the features. All they needed to do was update to a newer kernel (and hence graphics system, I believe WinMo only supported 16-bit color) and make a prettier UI, like what was in Photon. It could have been a hit had it come out on some decent hardware in 2008.

Instead they were stagnant, barely updating WM6 and held off until the end of 2010 to release WP7, which was extremely basic.

I'm very happy with my WP7, but it really is missing a lot of features that I had back in Windows Mobile. It is also much more restricted. It barely even does multi-tasking which WM did perfectly.

MFH said,

That's definitely Microsoft's fault, as both iOS and Android allow native code - so you can reuse your code for both platforms. WP7 is too focused on .NET and requires the developers to rewrite their programs…

That might change with Apollo ; we will probably see the same Windows Runtime, running on the same NT kernel, get ported to WP. Thus, there will be native code, without one of its main disadvantages for consumers (aka unrestricted platform access)

The NYT only forgot to say why Windows Phone hasn't got any attention from consumers... Because they are tired of Windows. About 95% of PCs have Windows. People want novelties, not that same old name they're obliged to use for decades.
Wake up Microsoft. You're screwed.

Luis Mazza said,
TBecause they are tired of Windows. About 95% of PCs have Windows. People want novelties,
And today, how is iPhone a novelty? It merely evolves from it's previous version without anything seriously new or groundbreaking. In fact at this point it's more a matter of catching up to what Android devices are featuring. And Talking about ANdroid.. not much news there either.. With it's hugely fragmented market it's hard to really innovate.

The only party which should be able to innovate now if they play it right (Humongous if..) will be MS. But seeing how they missed the ball on so many 'small' things in WP7 they will be busy fixing their oversights for the near future.

Their partnership with an innovative company like Nokia should be a pro, but personally I fear they will burn that down by not being able to catch up to what Nokia comes up with.

Nokia also rules a huge market where windows phone could make a huge dent, but if MS does not act quick enough it'll be too late and drag Nokia down with em..

You're thinking like an analyst, not a consumer, and they don't care about those deep analysis. I'm referring to novelties as "anything other than Windows-related".

MS not keeping up with Nokia?? MS knows how to support Dual/Multi-core CPUs, it's Nokia that don't know anything about that and are keeping that support back as MS wants Nokia to release the first DC-WP7 device…

Luis Mazza said,
The NYT only forgot to say why Windows Phone hasn't got any attention from consumers... Because they are tired of Windows. About 95% of PCs have Windows. People want novelties, not that same old name they're obliged to use for decades.
Wake up Microsoft. You're screwed.

Do you really think people buy Android just because they don't want "the same old name"?

Aethec said,

Do you really think people buy Android just because they don't want "the same old name"?

You mean "the same old name" = WINDOWS, than yes. Absolutely.
Nobody bears windows anymore. Names are much more important than you may think.

KingCrimson said,
So when did WP development begin, early 2009?

According to Wikipedia, yes I know, development may have started in 2004.

WelshBluebird said,
As others have said, this has been fairly well known for a while now.

This, nothing new here, left wondering wtf the point was of posting this article.

insanelyapple said,

Because whole Metro-thing on desktop PCs is just plain wrong.

Why do people link Operating Systems so strongly with the UI? When you start an OS from scratch the UI isn't the only freaking thing you care about ._.
A New OS is focused a LOT more on the core functionalities, kernel, programs etc.. I even doubt that they'll change the UI, if they were ever planning to recode Windows.
It's sad how much people just care about the UI these days. OK, I agree that it has a lot to do with user friendliness and stuff.. But people should take a better look at the REAL functionalities of an OS.

Yakuzing said,

That is incorrect.

Actually they did restart development with Vista. All the underlying technologies in Windows 7, the networking and audio to name only two of them, were introduced in Vista.

neo158 said,
Actually they did restart development with Vista. All the underlying technologies in Windows 7, the networking and audio to name only two of them, were introduced in Vista.

They restarted Vista's development, not Windows. There's a hell of a lot more to Windows than what was introduced in Vista/7.

Vista was a major milestone, which is hugely different to a restart from scratch.

neo158 said,

They already did with Vista so why would they do that again?


The difference with Vista was that, even after the dev reset, it was still based on a previous version of desktop Windows. I believe what they're trying to say is that WP7 was not based on a previous version of Windows Mobile.

Coi said,

Why do people link Operating Systems so strongly with the UI? When you start an OS from scratch the UI isn't the only freaking thing you care about ._.
A New OS is focused a LOT more on the core functionalities, kernel, programs etc.. I even doubt that they'll change the UI, if they were ever planning to recode Windows.
It's sad how much people just care about the UI these days. OK, I agree that it has a lot to do with user friendliness and stuff.. But people should take a better look at the REAL functionalities of an OS.

Why? Because youre working with GUI everyday, it is providing basic comfort for your work, doing tasks. Tablet UI for desktop machines is like i said before, plain wrong - just because they gain success with their phone, it doesnt mean that Metro will fit everywhere else. Dramatically change like getting rid of Start menu (which was greatest MS 'invention') in in the name of Metro success and faster release cycle is the stupidiest thing that i saw in IT world in my whole short life. Moreover, for now Windows 8 (DP and some beta leaks) looks like biggest inconsistency in design in whole MS history - tile UI mixed with "classic" Aero makes my eyes bleed. Of course Win 8 is still in working phase but faster release cycle (they switched around AFAIR 2007?) makes me believe that RC will not change dramatically.
I believe the truth is that they cant build new Windows from scratch and leave behind all current products, because they gain too big piece on marketshare and its far too late. For years they did of course good stuff like Aero, but its more like doing new origami figure from same, mussy chewed piece of paper.

Aethec said,

It's called Midori.

Well... not really, that's actually a project to write an OS (especially the kernel) with managed code among other things.

Brian Miller said,
They should do that for Windows.

History suggest that companies with the majority of the market share fail to innovate, until someone else forces them too

Coolicer said,

Well... not really, that's actually a project to write an OS (especially the kernel) with managed code among other things.


The question is if they will ever bring something like Singularity or Midori to the market. They returned to native code for the sake of performance and efficiency (see WinRT, see the "reborn" focus on C++). They've already once tried to use .NET for everything in Windows (Longhorn) and failed. Then they ditched the idea for Vista and 7; 8 will follow this route…

neo158 said,

Actually they did restart development with Vista. All the underlying technologies in Windows 7, the networking and audio to name only two of them, were introduced in Vista.


yep. Although Vista and XP appeared to look and function similarly, the underlying changes made to the OS was indeed an overhaul, thus all the compatibility problems with Vista.

MFH said,

The question is if they will ever bring something like Singularity or Midori to the market. They returned to native code for the sake of performance and efficiency (see WinRT, see the "reborn" focus on C++). They've already once tried to use .NET for everything in Windows (Longhorn) and failed. Then they ditched the idea for Vista and 7; 8 will follow this route…

Well everything that Microsoft Research is working on is pretty damn cool. Sure we won't see 1 on 1 implementations of stuff being worked at there, but just like any other far-fetched RnD project, some parts and stuff learned from the projects will flow back into the main products.

I think Midori is a research of the feasibility of a purely managed code kernel (well, you can never escape from writing some assembly in a kernel though).

Projects like Midori and Singularity give Microsoft (the Coding Gurus, Software Engineers, Software Architects, etc over there anyway, Microsoft just has too many suits working for me to talk about the company as a whole) insight on what is possible with their systems. WindowsNT has never been about 100% performance, WindowsNT is more about performing adequately while providing a stable and massive platform for corporations and businesses. If you want sheer performance Linux is tweakable to whatever hardware and task you're running. So in my opninion Microsoft Research is just exploring the boundaries of what course is possible to take with their platforms.

Man... how I wish to know all about the dark magic, top-secret, mystic projects they've got over at MS Research >.>

insanelyapple said,

Why? Because youre working with GUI everyday

People are acting like the GUI represents the whole OS. This is absolutely false, it's only PART of the OS and it can easily be changed / tweaked without needing to recode the whole freaking thing (like people think).

And for the Immersive UI part, they DID recode the UI, what do you think Metro is? And now people are complaining to recode it AGAIN while there's not even a Beta OS with the Immersive UI?
They might be complaining about the standard Desktop UI and taskbar, which they'll still see (!) on their desktops and laptops. Once again, I don't think that one needs a recode as they've been improving it with each new OS.

MFH said,

The question is if they will ever bring something like Singularity or Midori to the market. They returned to native code for the sake of performance and efficiency (see WinRT, see the "reborn" focus on C++). They've already once tried to use .NET for everything in Windows (Longhorn) and failed. Then they ditched the idea for Vista and 7; 8 will follow this route…

C++ performs better than .NET because .NET simply cannot be fast enough on Windows. Tons of .NET methods are in fact wrappers around Win32.
If you remove runtime systems and build managed code in the OS itelf (using a compiler such as Bartok which is almost exclusively written in managed code), you can get performance that is actually better than native code because everything is sandboxed. No context switches.

See "Abolish Runtime Systems!" a presentation PDF from Microsoft Research: http://research.microsoft.com/...rus/talks/vee06_keynote.pdf
The "Micro-Benchmarks" slide is pretty explicit.

FalseAgent said,

yep. Although Vista and XP appeared to look and function similarly, the underlying changes made to the OS was indeed an overhaul, thus all the compatibility problems with Vista.

Well, most of the compatibility issues were the result of program developers locking their programs to the major OS Version for some bizarre reason. Not a good practice clearly.

M_Lyons10 said,

Well, most of the compatibility issues were the result of program developers locking their programs to the major OS Version for some bizarre reason. Not a good practice clearly.

Yup. Which led to Win7 being NT 6.1, and Vista having _tons_ of junction points because idiot devs hardcoded paths such as Documents and Settings.

Aethec said,

See "Abolish Runtime Systems!" a presentation PDF from Microsoft Research: http://research.microsoft.com/...rus/talks/vee06_keynote.pdf
The "Micro-Benchmarks" slide is pretty explicit.

Interesting, though I'd like to see real-world benchmarks before fully believing their conclusions - we all know how easy it is to tweak results
BTW: one major problem with their system: heavy reliance on a proprietary language/platform which leads to decreased portability. Not to mention that all current managed platforms are deeply flawed in terms of resource management (just think about IDisposable and you know what I mean)…

MFH said,

Interesting, though I'd like to see real-world benchmarks before fully believing their conclusions - we all know how easy it is to tweak results
BTW: one major problem with their system: heavy reliance on a proprietary language/platform which leads to decreased portability. Not to mention that all current managed platforms are deeply flawed in terms of resource management (just think about IDisposable and you know what I mean)…

C# is not proprietary ; it is an open ISO standard, and anyone can implement it. Singularity isn't proprietary either, its source code is on CodePlex.
In fact, a fully managed OS is great because it can be used on all processor architectures with minimal work (porting the framework). IA64, despite all of its advantages, died because of native code compatibility.

Who actually uses the start menu in Windows 7? I have all my regular programs pinned to the start bar and use windows key shortcuts to open most other things. If I ever do open the start menu I never go beyond the first level main start menu.

insanelyapple said,

Why? Because youre working with GUI everyday, it is providing basic comfort for your work, doing tasks. Tablet UI for desktop machines is like i said before, plain wrong - just because they gain success with their phone, it doesnt mean that Metro will fit everywhere else. Dramatically change like getting rid of Start menu (which was greatest MS 'invention') in in the name of Metro success and faster release cycle is the stupidiest thing that i saw in IT world in my whole short life. Moreover, for now Windows 8 (DP and some beta leaks) looks like biggest inconsistency in design in whole MS history - tile UI mixed with "classic" Aero makes my eyes bleed. Of course Win 8 is still in working phase but faster release cycle (they switched around AFAIR 2007?) makes me believe that RC will not change dramatically.
I believe the truth is that they cant build new Windows from scratch and leave behind all current products, because they gain too big piece on marketshare and its far too late. For years they did of course good stuff like Aero, but its more like doing new origami figure from same, mussy chewed piece of paper.

neo158 said,

They already did with Vista so why would they do that again?

No they didn't. You are telling me they recoded Vista from the ground up? You really don't know much then. If it was built from scratch then a lot of apps would not run on that OS. It's a more bloated kernel with a lot of eye-candy added.

Microsoft should however take your advice and do that for after Windows 8. Windows 8 will not really take off IMO because a lot of companies tend to skip a generation. Most companies are migrating to Windows 7. But that's a different story all together.

neo158 said,

They already did with Vista so why would they do that again?

They did not "start from scratch", yes a lot of subsystems and stacks were rewritten. But it's just Windows NT still.

Aethec said,

C# is not proprietary ; it is an open ISO standard, and anyone can implement it.

Correct, but .NET is NOT! Only the first 2 versions of .NET were standardized. Have fun with them Matter of fact the language is way less important than the library… And as I already mentioned managed systems are still pretty bad in terms of resource management, hell I'd rather continue using C++ and relying on RAII than use this stupid IDisposable-paradigm…

UndergroundWire said,
No they didn't. You are telling me they recoded Vista from the ground up? You really don't know much then. If it was built from scratch then a lot of apps would not run on that OS. It's a more bloated kernel with a lot of eye-candy added.

NT 6 is certainly not bloated. Unless you consider security features such as integrity levels to be bloat.

Aethec said,

NT 6 is certainly not bloated. Unless you consider security features such as integrity levels to be bloat.

I get it, you are confusing bloated with bloatware. There is no doubt Windows got bloated. But the bloat is not crap. It could be security, eye-candy, usability features, etc.. The two should shouldn't be used synonymous.

Kirkburn said,

Probably because it works.

If it isn't broken don't fix it. But Microsoft really needs to rebuild the OS from scratch for the tablet. Windows 8 is just a two-headed monster with just a lot of bloat for a tablet. Apple really did the job right on this one. They new how to differentiate a "ultra" mobile OS from a desktop OS.

Brian Miller said,

Windows 7 still has code from Windows 3.1.


And NO other OS has code from older versions!!!!
D:

That would have made a good article...three years ago.
Just because the NYT says it now doesn't mean it's news.

(btw, that NYT article lacks a critical fact: Microsoft was already preparing a WM evolution codenamed Photon, but ditched it)

Maysky said,
Three years ago, Microsoft would not had admitted it, with time passed, they've opened up a bit.

Uhh, when they first showed off Windows Phone 7 at MWC over a year ago they said that they started from scratch and that's why some key things were missing from the 7.0 (really v1.0) of the product at the time. This really isn't news now.

GP007 said,

Uhh, when they first showed off Windows Phone 7 at MWC over a year ago they said that they started from scratch and that's why some key things were missing from the 7.0 (really v1.0) of the product at the time. This really isn't news now.

Exactly... I'm wondering why this is suddenly news when it was publicly announced over a year ago...

GP007 said,

Uhh, when they first showed off Windows Phone 7 at MWC over a year ago they said that they started from scratch and that's why some key things were missing from the 7.0 (really v1.0) of the product at the time. This really isn't news now.


^this.

Aethec said,
That would have made a good article...three years ago.
Just because the NYT says it now doesn't mean it's news.

(btw, that NYT article lacks a critical fact: Microsoft was already preparing a WM evolution codenamed Photon, but ditched it)

I couldn't agree more. Microsoft seems to drag the line a bit on some of their product lines, often losing out on marketshare, and this is a prime example.

Maysky said,
Three years ago, Microsoft would not had admitted it, with time passed, they've opened up a bit.

As GP007 said, Microsoft has been very open about this... This is a very odd "news" story...

GP007 said,

Uhh, when they first showed off Windows Phone 7 at MWC over a year ago they said that they started from scratch and that's why some key things were missing from the 7.0 (really v1.0) of the product at the time. This really isn't news now.

They pretty much had to show it off early. That is a big conference. What were they going to do show it off in MWC 2011?