Michael Tsang - What language/tool did you use to write the Tablet PC's drivers?

Sign in to queue

Description

Michael Tsang is a software design engineer on the Tablet PC team. What does that mean? He's the guy who wrote the driver software for the digitizer. He also wrote the software that handles what happens when you push on the buttons on the Tablet PC (like the ones that make your page scroll up or down).

Translation: he isn't a C# coder.

Find out what he uses to code and why he can't use .NET to do his work.

Embed

Download

Download this episode

The Discussion

  • User profile image
    Charles
    There will come a day when the CLR will find it's way into kernel mode. It's inevitable.

    Charles
  • User profile image
    sbc
    I wonder what kind of PC's we will have then? Average PC: P6 XeonItanium Ultra 8GHz with 4GB RAM, HDDVD player and 1TB hard drive. Quite a long time away I think.
  • User profile image
    Charles
    That's quite a system, SBC. Smiley 

    It certainly will require advances in hardware as well as software, but I bet it won't require quite the horsepower of the system you envision. Though, such a system could very well exist in the next ten years or so which is about how long it will take for the advent of a managed kernel. It's certainly possible and some Microsoft minds are thinking about it as we type.

    Charles
  • User profile image
    Senkwe Chanda
    Makes me wonder how the Linux Kernel devs would react. Dismiss it as a bad idea (and it might well be), or go the same route. If they go the same route, which managed environment do they choose?

    Interesting times Smiley
  • User profile image
    Charles

    It could be a bad idea (I don't think so), but that won't be able to be determined until it actually happens. 

    The hypothesis here goes something like this:

    The vector of software abstraction in advanced operating systems will continue pointing in the positive direction as time increases (meaning more and more managed code can accomplish more and more tasks that once required writing unmanaged code). The average speed with which these new concepts are made concrete will not decrease over time. Win32 to WinFX is a good example of this type of advancement in software abstraction in Windows.

    I see no good reason why this abstraction cannot be applied to the kernel (except that it will be really hard to do and will require innovation on the hardware level). Things like GC memory management and guaranteed type safety in the kernel will increase system stability and reliability without incurring too severe of a performance penalty (though of course there will be a cost. Nothing is free...).

    I'm looking forward to seeing how deep the CLR will go into Windows in the future (it's pretty deep in Longhorn, but there's still plenty of room for expansion).




    Charles


  • User profile image
    barlo_mung
    I think there will be more of a convergence here, at first anyway.  WDF will have support for user mode drivers that I believe can be managed code. So we may see more functionality move up out of the kernel into user space leaving the kernel space the domain of the highly skilled people like Mike.
  • User profile image
    W3bbo
    Charles wrote:
    There will come a day when the CLR will find it's way into kernel mode. It's inevitable.

    Charles


    Sorry for digging up this year and a half old thread (it appeared on the "Featured" bit), but I thought Wacom developed the digitizer, not Microsoft.

Add Your 2 Cents