As WARM (Windows-ARM) reportedly takes the stage in Las Vegas tomorrow, some thoughts:

  • This is “big” Windows, not yet another repackaging of Windows CE.  Remember Windows NT got its start supporting multiple CPU architectures.
  • This has huge implications for Windows Phone’s future.  It makes no sense to have two separate operating systems and application ecosystems for increasingly overlapping touch devices (phones and tablets).  It sucks for customers, developers, OEMs and is terrible for Microsoft economically to have to build and support parallel operating systems.  And Windows Phone’s road is profitability is hard to imagine.  Even if you assumed a wild leap to 20% market share, at <=$10/unit, it isn’t going to pay for the 3,000+ people working on it, never mind the marketing spend and OEM “incentives”, any time soon.  Windows Phone 8 probably is a configuration of big Windows on ARM which lets that team focus on the phone experience and not have to build an operating system top to bottom.
  • This also explains the demise of Courier – a third operating system in the mix would be exponentially worse.  Presumably the Courier application experience is being implemented on big Windows as the shell for Windows tablets.
  • While iOS is Apple’s branded operating system for touch devices, it shares the same underlying kernel, tool chains, etc. with Mac OSX.  Microsoft aspires to have a single, modular operating system that can be factored appropriately for the increasing variety of form factors.  Better modularization will also help power efficiency from a software perspective.  Expect new configurations of big Windows for TVs, settop boxes, etc.
  • In theory Windows apps can be recompiled for ARM, but in reality they all need new user interfaces for the touch world.  So much for the vaunted “applications barrier to entry”.
  • Meanwhile, the modest traction Microsoft is making with application developers for Windows Phone 7 is at risk as it is not clear whether the Windows Phone application model will be supported in the future or whether something new will be introduced.  History suggests the big Windows team will have opinions on the application model.
  • The ARM support won’t show up until Windows 8 (presumed to be 2012), which is an awfully long time to wait.  The incredibly late to materialize Windows 7-based tablets look like sacrificial offerings.  Meanwhile, analysts variously estimate Apple ships between 30 and 50 million iPads this year.  And we’ll see how whether Android 3.0 is as successful with tablets as it was with smartphones, with devices hitting the shelves shortly.  There is a huge difference between being number two and number three in  market (and I guess I should mention RIMM and WebOS for completeness and the possibility Microsoft could be number five in this market).  Microsoft might consider stopping spotting multiple competitors multi-year leads in some of these markets.  But maybe the company just likes a good challenge.
  • Needless to say, Microsoft is in a tough position.  Getting to a single operating system and single application model is desirable for the long term, but the degree of difficulty to get there is incredibly high being a year or more from shipping product, having a full slate (yuck, yuck) of competitors in the market and potentially Osborning the current Windows Phone along the way.
  • But it could be worse – you could be Intel.  Microsoft porting to ARM is a serious indictment of Intel’s power efficiency roadmap.  Historically, Intel-Microsoft executive meetings have had colorful moments and I’d pay to see video of some of the recent ones.  I do expect Microsoft to take the high road and throw Intel a conciliatory bone or two, deeming the next generation of Atom chips to be “pretty good (for you guys…)”.  And while client-focused, this move also improves Microsoft’s options for supporting ARM-based servers in the future, making this a double-barreled nightmare for Intel.  But at least they control their own destiny with that MeToo, er, MeeGo operating system.

Will be fun watching to see how Redmond plays this one.

4 responses

  1. dead on, Charles. Of course, I expect no less… 😉

  2. Charles, I love this one:

    History suggests the big Windows team will have opinions on the application model.

    Ya think?

  3. I guess, I’ll be the discenting opinion.

    Big windows may be ok on a tablet, but it will take a long time to reach the power requirements of a phone. So although the scenario of the Big Windows kernel on a phone may play out in the long run (4+ years) it will not happen soon, and certainly not for the next 2 generations of Windows Phone 7. Also there is no issue with application compatibility for developers to the Window Phone 7 platform since that platform is .NET, it will still run on “big windows”, it is just a question of building support for both into the OS (which should not be difficult since they will be very close).

    Also Windows Phone 7 is NOT another repackaging of Windows CE, it is a ground up new operating system. Of course since it supports .NET and that is the application model, the operating system is beside the point for developers.

    This iOS/OSX argument isn’t really relevant, it is the application model that needs consistency not the operating system kernel. The team developing the OS is relatively small for Windows Phone 7, and so is not very relevant to the economic argument.

    Also remember that Big Windows does today run on some small platforms. The kernel is XBOX 360 is a stripped down version of Big Windows, in a platforms with only 512MB of RAM. The problem though is the XBOX 360 can consume a lot of power, unlike the phone or tablet, so there are additional challenges.

    I think the argument that apps must be re-authored isn’t correct either. They will certainly be better with reauthoring, but a UI that understands coverting the inputs of a tablet and multi-touch to existing Windows UI model can mitigate this mostly.

    I think Windows future is bright on all 3 platforms, and the bigger point is Microsoft is never going to stop. There is no point where they will not be on all 3 of these platforms and improving their game. That alone should give developers the confidence to invest.

    A developer trying to develop across platforms still will only have 1 choice for a consistent framework across platforms, and that is Windows and .NET. So as labor becomes the most important economic input to development, as hardware drops to near zero cost, the cheapest platform for developers will be Windows.

  4. Aaron,

    The question is if I want to build a touch app, what application model should I write to and what Windows devices will it run on when? Just saying it is Windows/.NET doesn’t answer that question. It isn’t clear .NET will play any role in Microsoft’s tablet strategy.

    Meanwhile I can do Android or iOS apps and easily target both smartphones and tablets.

Get Updates By Email
%d bloggers like this: