Coffeehouse Thread

29 posts

Forum Read Only

This forum has been made read only by the site admins. No new threads or comments can be added.

Is There a Need for a Free Windows Kernel?

Back to Forum: Coffeehouse
  • User profile image
    giovanni

    I recently purchased a couple of managed switches for our small office and I discovered with horror I had to configure them with a good old VT100! Lets just say that in a futuristic IT environment, I have a hard time appreciating the tool.

    As I am slowly learning PowerShell, I find it far superior to any Linux/UNIX shell (except for the terminal itself) I have been using; wouldn't it be great if it was available on network equipment? In other words: could Microsoft license for free the Windows kernel with some tools for use in such devices in a similar way as Microsoft supports the .NET Micro Framework?

    This might sound a little utopic, but in a world where Microsoft has to compete more and more with open and free software this would also have the advantage of letting people play with the OS and expand it further. It is undeniable that the availability of the Linux kernel is one of the main reasons it is so widely spread today and it is easy to find people with good expertise on it.

    Alternatively, would it be possible to port some of the tools like PowerShell to Linux? This would not have the advantage of spreading the OS, but it would probably be a good compromise to introduce people to the tool...

  • User profile image
    fanbaby

    Well... where do I start...

    Maybe just this: "Unix, just when you think you're out, they PULL YOU RIGHT BACK IN"

    I know PowerShell is the creation of a guy or two in isolation, but it might have very well been a product of Gates-Ballmer's Microsoft: get some very decent developers, copy an existing product, open-source or not, learn from design mistakes and make an arguably better product.

    PowerShell might be a better shell, and also open-source, but my advice to you is: no cross-platform, no buy. JUST LEARN BASH.

    It's a typical MS ghetto technology.

  • User profile image
    giovanni

    Except I am way more productive in PowerShell than I have ever been in Bash... My point is this: PowerShell is a great product for IT, what strategies could make it more ubiquitous?

    Product B is worse than product A, therefore stick to product B does not sound like good advise to me.

  • User profile image
    Bass

    Bash is "good enough" that I don't think any Linux sysadmins are going to migrate to Windows for PowerShell. So I don't think PowerShell is really much of a competitive advantage.

    What this means is I don't think it would be risky to Microsoft to open source it. In fact by open sourcing it, it might become bigger on non-Microsoft OSes and reverse misconceptions that Microsoft doesn't have any idea how to do shells, thus improving PR in an area where Microsoft has little. And might interest those same Linux sysadmins in other Microsoft products.

    Basically what I am saying is I find it hard to find any reason NOT to open source PowerShell.

    As so far in the Windows kernel, I don't think it's going to be ever competitive with Linux in networking equipment because Linux has been optimized for this purpose for years. So I imagine it would be one of those things that is open sourced and promptly ignored? Who knows.

  • User profile image
    giovanni

    , Bass wrote

    What this means is I don't think it would be risky to Microsoft to open source it. In fact by open sourcing it, it might become bigger on non-Microsoft OSes and reverse misconceptions that Microsoft doesn't have any idea how to do shells, thus improving PR in an area where Microsoft has little. And might interest those same Linux sysadmins in other Microsoft products.

    Yes, this is part of my thinking. I think the coexistence of Windows and Linux is here to stay, therefore having a corss-platform shell like PowerShell could only be positive both ways, but has some good potential particularly for Microsoft.

  • User profile image
    Bass

    , giovanni wrote

    *snip*

    Yes, this is part of my thinking. I think the coexistence of Windows and Linux is here to stay, therefore having a corss-platform shell like PowerShell could only be positive both ways, but has some good potential for Microsoft.

    Yep. I think open sourcing PowerShell would be pretty smart. I also noticed that a lot in the PoSh community seem to want this too - if it is available on more systems their investment in learning the shell becomes more valuable. This is true also for people who have yet to use PowerShell, it sweetens the deal of learning it.

    The Windows open source thing though:

    • If Microsoft open sources only enough of Windows that it wouldn't be a risky business move (ie. only the very core kernel), the end result wouldn't be very useful.
    • If Microsoft open sources large parts of Windows (eg: the Win32 API), it would be a very risky business move because Windows is a big money maker for Microsoft in general. The ability to run Windows software is a huge competitive advantage of Windows. So anything having that feature could be very risky.

    Not that I have anything against Microsoft open sourcing anything, but I find this one a little tougher to justify from a business perspective. Smiley

  • User profile image
    bondsbw

    Google led me to this:  http://pash.sourceforge.net/

    But I haven't used it to know how well it works.

  • User profile image
    giovanni

    @Bass: I think at some point they will have to do something if they still want to be a software company (which they probably don't have to if they become a devices and services one, but still...): the open source model is clearly a winning one when you look at a number of devices and not just network switches.

    With a smaller and smaller Windows kernel (Windows Phone runs it already), I don't see why Microsoft should keep Windows CE on forever. I think at some point it will make sense to only keep and maintain one. However, they need a kernel for embedded devices either for third party manufacturers or for themselves (think about the potential for wearable electronics or for modular components like Surface keyboards...). The most obvious way for me would be to open source at least enough of the kernel to be used in such devices.

    As you said, this would spark a lot of interest in other Windows products and would create a much bigger community of developers and sysadmins.

  • User profile image
    fanbaby

    , giovanni wrote

    Product B is worse than product A, therefore stick to product B does not sound like good advise to me.

    Plan 9 is much better then unix, yet it hasn't a chance to replace it. Check their faq.

  • User profile image
    Bass

    Yeah but they still make tons of money on selling Windows. They won't open source anything that would even remotely threaten that revenue source. Without the Win32 API a lot of the use case for Windows falls apart, because really nobody is like "I'm going to use Windows for their kernel architecture design". It's totally vanilla stuff. So an open Windows kernel, I don't know what it will accomplish?

    So to sweeten the deal they'd have to open source things that do in fact provide Windows with the competitive advantage, and that could threaten their business. At least that's the way I'm seeing it.

    To be fair, the competitive value of the Windows API is in a decline. Microsoft is moving away from that sort of thing to become a "device and services" company. So maybe in a few years, especially if they truly turn into a device and services company (they are at best, at a transition stage right now). They are still a software company right now if you read the financials. Open sourcing Windows might make sense for a "device and services company", but not for a software company. But we'll see. On the order of decades, it is something that could happen.

  • User profile image
    Bass

    It's worth noting that even companies that have strong relationships with FOSS, they never open source their "crown jewels" so to speak. Google won't open source the software that drives their search engines and what not. They'll open source utility libraries they use in the search engine that aren't exactly huge competitive advantages, eg. protocol buffers. But even then, only when they think it could be useful to the FOSS community and serves some business purpose to become open. If something is open sourced and not used, it serves no purpose for the company to FOSS' it, and there is overhead in preparing something to be FOSS - so it's not entirely free to do.

    But, if it is used they gain:

    • PR
    • Often, free engineering and ideas for improvements that people in the company might have not had

    And that could be very valuable. That only applies if it is used of course.

  • User profile image
    giovanni

    @Bass: I agree with you, but:

    • Windows is not their biggest source of income
    • If they become a Devices and Services company that could change a lot of dynamics
    • There might be other way to protect Windows as a platform

    I am no historian, but it seems to me that Windows was born from the need for an OS because Microsoft wrote applications for the Mac and they wanted a larger base than what it was possible if Apple did not license its OS to other manufacturers. Gates bought DOS because he wanted to write applications that run on an OS installed on more than just on the hardware of a single manufacturer.

    I would bet that if today Microsoft needed an OS the obvious choice would be another one...

  • User profile image
    fanbaby

    @Bass: Open sourcing windows or dotnet wouldn't change a thing. See mono's case. I think that most of mono's users are MSFT developers.

    But you know what will? When the new management will reverse their sue-happy policy. That, I think is Ballmer (and Gates) true legacy. The legacy of an MBA who truly believes that Microsoft invented computing. Cannot make an OS kernel or a browser without stepping on 287 of Microsoft's patents. Bye-bye Ballmer. Will not miss you as a patent abuser and a bully.

  • User profile image
    giovanni

    , giovanni wrote

    • There might be other way to protect Windows as a platform

    Le me state this another way: they need to make Windows more readily available for embedded devices. Open sourcing it would be one way, there are probably others.

  • User profile image
    Bass

    The problem is they are still stuck with their legacy as a software company that makes most revenue from licensing. It's a tough nut to crack.

    They still have to be protective of their "legacy" as a software company because that is STILL where most of the money that pays the bills comes from. So while Google can be like, fine lets open source Android. Why not? We make all our money from advertisements and services anyway.

    Microsoft can't do that, because they still make most of their money from conventional licensing. If they open source too much, they can threaten that revenue source. Google doesn't have that problem. Microsoft can't survive as a company without their traditional revenue sources, at least today. Again, Google doesn't have that problem. That's what they want to change though.

    Google doesn't any of these problems, so they can take risks that Microsoft wouldn't even consider. And they did.

    Microsoft weren't in the right place as a company (ie. they were still a software company, when they really needed to be a device and services company) and that resulted in Android dominating Microsoft in what really is its own game (platforms). To be fair, I don't think Microsoft was prepared to have "competition" in the first place. Their development models and business processes feel designed towards a company who is used to have a monopoly in client OSes, not one that needs to rapidly react to competition. As they say, success is a disease.

    Considering Linux and Android position in the embedded/mobile market, I don't see this as reversible. As far as Microsoft on embedded, I feel like that's pretty much a lost cause at this point. Even if they open source Windows, they are now competing with other open source that is more established. So, what I am saying is it would be a too little, too late. They've lost that market I think, much like IBM lost clients in the 90s.

  • User profile image
    giovanni

    @Bass: again, I tend to agree with you, but I don't think everyone likes the advertising and services model.

    Microsoft needs to make some difficult decisions to transition to a devices and services company and I don't think any of them would be impossible given the right compromise and the right time.

    Not long ago Blackbarry was dominating the smartphone market, then it was Apple, now it is Android. Things change pretty quickly in this industry and it does not look to me that Microsoft will disappear tomorrow...

  • User profile image
    ZippyV

    Never heard about Windows embedded?

  • User profile image
    Blue Ink

    , ZippyV wrote

    Never heard about Windows embedded?

    The problem is, how much are you willing to pay for an OS given that you plan to sell your device for $50?

    The other fundamental question is how much of an OS you really need, or if you need one at all.

    That's where Microsoft could really make a difference: tools in embedded space are best described as primitive, and often enough their choice is dictated more by what is free than by what works best. I bet lots of developers would be more than happy to pay for VS if it could be used for firmware work (VisualGDB kind of works, but it's silly that you have to use GCC in the backend).

Conversation locked

This conversation has been locked by the site admins. No new comments can be made.