01.18.09
Why Microsoft Wants to Put .NET/Mono in Devices
Novell & Miguel de Icaza help Microsoft
SOME people may be wondering why Microsoft’s partners want Mono in LiMo and why Miguel de Icaza is so keen on putting Mono inside Android [1, 2]. Microsoft Windows CE/Mobile is not doing well in this area, so the company wants its Intellectual Monopolies to ‘infect’ competitors’ devices. Today we’ll show evidence from Microsoft, in the form of E-mails that came through antitrust litigation.
“I saw that internally inside Microsoft many times when I was told to stay away from supporting Mono in public. They reserve the right to sue”
–Robert Scoble, former Microsoft evangelist
In today’s evidence, Microsoft discusses the failure of Windows CE and the last resort — a need to “license” from competing operating systems, as we shall show at the bottom.
In a world or a nation where software patents are seen as legitimate, ownership need not be physical alone, so there is wiggling room. How to do all that? Formats, programs (like codecs) and .NET, even DRM.
As a timely plug from the news, Ryan Paul is boosting Mono again. It’s ‘infecting’ iPhone and Wii, two devices that Microsoft is very much afraid of.
According to Novell’s lead Mono developer, Miguel de Icaza, several applications in Apple’s App Store are powered by Mono. This might come as a bit of a surprise to those familiar with Apple’s highly restrictive application inclusion policies, because the company strictly prohibits developers from using interpreted languages and third-party runtime environments—a constraint that largely rules out technologies like .NET and Java.
Many comments appear in the accompanying OS News pointer. Mono is very controversial for various reasons.
Arstechnica reports that Mono, an open source implementation of .NET runtime, is bringing Microsoft’s development technologies to some unexpected places, including the iPhone, Android, and the Wii.
Mono continues to evolve and sneak into areas that are legal minefields, by admission from the most stubborn defenders of Mono, such as Dan O’Brian. From Heise:
Version 2.2 of Mono, the open source alternative to Microsoft’s .NET framework, is now available.
Evidence of Microsoft’s plans for Mono/.NET we have covered using antitrust material before (E-mails from inside Microsoft). But today we move on and look at Comes vs. Microsoft Exhibit px07010 [PDF], which we translated entirely by hand (it’s too poorly scanned). We address this reverse chronologically, just like in the exhibit.
In this “Highly Confidential” set of E-mails from 2001, Microsoft’s Alex Limberis writes:
- Customer visit evidence that 50% plus is Linux
- CE’s POR market data pointing to only less than 10% market share
- Support from MSTV agreeing we need to support all OSes
He is talking about set-top boxes. For context, Alex Limberis was “Director of Wireless, Broadband and Consumer Electronics” at Microsoft between 2000 and 2004.
Microsoft’s Amir Majidimehr, who left Microsoft a year ago, writes:
In Korea, clearly the format is dragging the OS in. But elsewhere in the world, where we having nothing close to 90% (or even 90%) marketshare, the OS is selected first.
Microsoft wants to control the format (or API) and it uses Korea as an example. Just watch the Korea/ActiveX situation [1, 2, 3].
I think the situation is rather simple (although the decision is not). When we asked for permission to take our format everywhere 2.5 years ago, no one blinked. JimAll himself said it was just fine to go on embedded Linux. Of course, this was at a time when people though we brought no compelling value.
Control the format, control the minds.
Now, contrast this with the situation in Korea where WMT has a 90% share. Last time I was there, I think they said they have something like a dozen WinCE stb wins, all because of WMT. You better not say anything about supporting Linux or the ESG guys escort you out to the airport!
Eureka moment:
So, the story at both extremes is very clear. When we had no value, we could be everywhere. And where we have ultimate value, then we must be on our own platform. The question is, when do you make the switch?
Integration and dependency:
[Alex Limberis] To be a format means render everywhere. The key to OS success is to create the value proposition from a compelling combination of ease of use, time to market and features of which WM, IE, real time kernel, IP stack, etc. etc. is part of the package.
Kind of like .NET versus Mono… and Moonlight with Microsoft’s proprietary codecs. To quote further:
With nearly zero design wins for video, and strong threat of MPEG-4, I would say this is no time to be picky. We must establish our base as we have done with audio. The stronger we become, the more value we can provide to the other assets of the company. We need to be the “icing on the cake and not the cake itself” when it comes to WinCE! This is not what happened PocketPC.
Linux-powered devices like TiVo are then mentioned:
[Alex Limberis] In the case of audio, we were not facing a “thick” os. The CE team never felt threatened by the OS as it were running on RIO. They are threatened by the OS running on a Moto, Replay TV, Ti Vo etc. that could support our format. This is our internal challenge.
Then it’s back to media formats and their power as a “standard” (which is controlled by Microsoft through bundling or added value… even software patents).
I always come back to; we are FM radio, or NTSC TV we must render on everything regardless of what shape, form or OS choice the gizmo is made from. CE must win on the total value of proposition of ease of use and integration of technologies including WM. In total power cost to deploy.
Linux and the GPL are mentioned as an “issue” (because Microsoft hates them both, of course).
His expressed sentiment seems to be the blocker for us moving forward and not the Linux GPL issue.
Here is why Mono may merely be a path towards Windows and Visual Studio in the long term:
I don’t know what other argument to make other than we must have 100% of the nodes available to us to be a successful format, and no matter how you slice, dice, justify and segment CE doesn’t have anywhere near 100%. Then we get into the belief that having WM in account helps move them to a Win OS in the future.
It then moves on to .NET:
The .net argument doesn’t cut it. Doesn’t .net need to run on everything also?
Here is a key part which is very telling:
2. As I mentioned on the call, WMT is one of the main distinguishing features of WinCE. To offer it on Linux (starting with STBs and then extending to other devices as well) robs us of our competitive edge.
3. While you mentioned that you will start by offering only the codecs and Udl on Linux today, I am very worried that a year from now, you will need to bow to customer demand and offer WMP and DRM as well at that stage. And by that time, the horse may have bolted too far for us to lock the gates.
4. There was mention of OEMs such as Pace, Moto, Nokia moving to Linux today. It is true that many OEMs are playing around with Linux today (since it is the easiest for developers to kick tires on)….however, most of the same OEMs mentioned above are moving to a WinCE platform as we discuss this. And the move in most cases has been driven by their need for WMT – if you take that away, we lose the battle before it begins.We understand the need for DMD to proliferate the format … however, if it is at the expense of our embedded OSs, I fear that it can hurt us in the long run. Especially when you consider that part of enabling the .NET vision is to embed our OSs in devices of all forms going forward – hence our apprehension.
Here is the bit about “we must license onto the competing OSes” in order to compensate for Windows’ failure in this space.
I guess the net of it from DMD’s perspective is only 6% of the world is going to be CE based we must license onto the competing OSes.
So there. They want to pollute the competition with their own technology. They see this as a gaining strategy. █
“As many of you may know, we’ve actually kind of broadened the product portfolio of Visual Studio, targeting all the way from the low end with students and hobbyists, kind of competitive in that Linux space, making sure that every developer has a copy of .NET and is trained in writing .NET solutions. [...] I think it will really help us in our competition with open source.”
–Eric Rudder, Senior Vice President, Microsoft
Appendix: Comes vs. Microsoft – exhibit px07010, as text
From: Alex Limberis
Sent: Saturday, October 27, 2001 9:25 PM
To: Amir Majidimehr; Will Poole; Kurt Buecheler
Subject: RE: STB ms data
I want to paint a clear picture for us. If you really do add value by holding on for MS OS’s then we shouldn’t sell ourselves short.
Question to Will. How much data do we need to make the point?
We have
- Customer visit evidence that 50% plus is Linux
- CE’s POR market data pointing to only less than 10% market share
- Support from MSTV agreeing we need to support all OSes
—– Original Message ——
From: Amir Majidimehr
Sent: Saturday, October 27, 2001 3:34PM
To: Alex Limberis; Will Poole; Kurt Buecheler
Subject: RE: STB ms data
In Korea, clearly the format is dragging the OS in. But elsewhere in the world, where we having nothing close to 90% (or even 90%) marketshare, the OS is selected first.
Amir
—– Original Message ——
From: Alex Limberis
Sent: Friday, October 26, 2001 11:17 PM
To: Amir Majidimehr; Will Poole; Kurt Buecheler
Subject: RE: STB ms data
I want to both quantify the positive, i.e. the Korean example as well as the current situation appearing in other markets where it seems the OS/processor decision was made first and the format is made second. Do we have more concrete data, i.e. where is the 90% WMT share? Other comment in line
—– Original Message ——
From: Amir Majidimehr
Sent: Friday, October 26, 2001 11:04PM
To: Alex Limberis; Will Poole; Kurt Buecheler
Subject: RE: STB ms data
I think the situation is rather simple (although the decision is not). When we asked for permission to take our format everywhere 2.5 years ago, no one blinked. JimAll himself said it was just fine to go on embedded Linux. Of course, this was at a time when people
though we brought no compelling value.
Now, contrast this with the situation in Korea where WMT has a 90% share. Last time I was there, I think they said they have something like a dozen WinCE stb wins, all because of WMT. You better not say anything about supporting Linux or the ESG guys escort you out to the airport!
So, the story at both extremes is very clear. When we had no value, we could be everywhere. And where we have ultimate value, then we must be on our own platform. The question is, when do you make the switch?
[Alex Limberis] To be a format means render everywhere. The key to OS success is to create the value proposition from a compelling combination of ease of use, time to market and features of which WM, IE, real time kernel, IP stack, etc. etc. is part of the package.
With nearly zero design wins for video, and strong threat of MPEG-4, I would say this is no time to be picky. We must establish our base as we have done with audio. The stronger we become, the more value we can provide to the other assets of the company. We need to be the “icing on the cake and not the cake itself” when it comes to WinCE! This is not what happened PocketPC.
[Alex Limberis] In the case of audio, we were not facing a “thick” os. The CE team never felt threatened by the OS as it were running on RIO. They are threatened by the OS running on a Moto, Replay TV, Ti Vo etc. that could support our format. This is our internal challenge.
Amir
—– Original Message ——
From: Alex Limberis
Sent: Friday, October 26, 2001 10:32 PM
To: Will Poole; Kurt Buecheler; Amir Majidimehr
Subject: RE: STB ms data
If WMT is the Holy grail, then I’d hate to take it away from them, and we should feel damn proud that we have created such a compelling value that they will switch OSes just to get WMT. Great job all, we have won!
However, I’m not convinced that our position is that strong.
I always come back to; we are FM radio, or NTSC TV we must render on everything regardless of what shape, form or OS choice the gizmo is made from. CE must win on the total value of proposition of ease of use and integration of technologies including WM. In total power cost to deploy.
His expressed sentiment seems to be the blocker for us moving forward and not the Linux GPL issue. I don’t know what other argument to make other than we must have 100% of the nodes available to us to be a successful format, and no matter how you slice, dice, justify and segment CE doesn’t have anywhere near 100%. Then we get into the belief that having WM in account helps move them to a Win OS in the future. I believe that, but he doesn’t.
Second, he has what I believe is the misunderstanding of the value of the OS vs. format. He believes they will pick the format and then figure out what OS is needed to support the format. Where, I believe that the OS is critical to the box decision. Motorola chose the PowerPC based on the fact that they own part of the processor and then created a box with compelling applications.
The .net argument doesn’t cut it. Doesn’t .net need to run on everything also?
Thoughts, before I jump on them?
—– Original Message ——
From: Srivats Srinivasan
Sent: Friday, October 26, 2001 5:14 PM
To: Alex Limberis
Cc: Takeshi Numoto
Subject: RE: STB ms data
Alex -
thanks for inviting us on the call today. First to answer your question below – this is IDC 2001 data and while WinCE share is shown as fairly small (Linux is even smaller)… and we expect the WinCE share to grow faster now that we’re actually putting resources behind this effort at our end.
Based on the conversation then, we have a few concerns that we need to discuss/address -
1. based on the feedback that we’ve got from customers, the most likely scenario appeas to be that the NetOp or Telco makes a format decision and then dictates to the OEM what box to build to meet those needs, i.e. the OS decision is minor and is not a casual factor in the process.
2. As I mentioned on the call, WMT is one of the main distinguishing features of WinCE. To offer it on Linux (starting with STBs and then extending to other devices as well) robs us of our competitive edge.
3. While you mentioned that you will start by offering only the codecs and Udl on Linux today, I am very worried that a year from now, you will need to bow to customer demand and offer WMP and DRM as well at that stage. And by that time, the horse may have bolted too far for us to lock the gates.
4. There was mention of OEMs such as Pace, Moto, Nokia moving to Linux today. It is true that many OEMs are playing around with Linux today (since it is the easiest for developers to kick tires on)….however, most of the same OEMs mentioned above are moving to a WinCE platform as we discuss this. And the move in most cases has been driven by their need for WMT – if you take that away, we lose the battle before it begins.
We understand the need for DMD to proliferate the format … however, if it is at the expense of our embedded OSs, I fear that it can hurt us in the long run. Especially when you consider that part of enabling the .NET vision is to embed our OSs in devices of all forms going forward – hence our apprehension.
—– Original Message ——
From: Alex Limberis
Sent: Friday, October 26, 2001 1:19 PM
To: Srivats Srinivasan; Phil Corman; Den Polling
Subject: RE: STB ms data
Thanks, what was the source of this data? I guess the net of it from DMD’s perspective is only 6% of the world is going to be CE based we must license onto the competing OSes.
—– Original Message ——
From: Srivats Srinivasan
Sent: Friday, October 26, 2001 11:17 AM
To: Alex Limberis, Phil Corman; Den Polling
Subject: STB ms data
Srivats Srinivasan
EAPG
x53975
[end of appendix]
“There is a substantive effort in open source to bring such an implementation of .Net to market, known as Mono and being driven by Novell, and one of the attributes of the agreement we made with Novell is that the intellectual property associated with that is available to Novell customers.”
–Bob Muglia, Microsoft President
Content is available under CC-BY-SA