Microsoft postpones PDC

Mary Jo Foley (she’s been covering Microsoft for a long time) has the news: Microsoft has postponed the PDC that it had planned for later this year.

The PDC stands for “Professional Developer’s Conference.” It happens only when Microsoft knows it’ll have a major new platform to announce. Usually a new version of Windows or a new Internet strategy.

So, this means a couple of things: no new Windows and no major new Internet strategy this year.

Contrast this to Google who is holding a huge developer day next week (it sold out, so I won’t even bother linking to it). Or Facebook, who held a big developer-centric shindig today.

Some other things I’m hearing about the next version of Windows? There still is a ban on .NET code in core parts of Windows. They aren’t getting enough performance yet from .NET to include code written in it inside major parts of Windows. This is a bummer, because .NET is a lot easier to write than C++ and letting Microsoft’s developers write .NET code for Windows would unleash a bunch of innovation.

The person who told me this (who works at Microsoft) told me .NET still takes too long to startup and load into memory and because Windows is now being compared to OSX they can’t afford to ship components that would slow down Windows.

Before every MVP jumps me in the alley yes, I know the .NET runtimes ship with Vista. But almost no Vista code was written in .NET (if any, actually). Microsoft tries to keep this secret because they know it gives a black eye to .NET. After all, if Microsoft is unwilling to use it to develop Windows or Office, why should the rest of us base our life on it? Easy, it’s a lot more productive for the rest of us to write code in .NET and now Silverlight, which uses .NET’s compiler and part of its framework at heart, than to fall back to C++. Pick the right tool for the job and all that.

It also means that Ray Ozzie’s team probably doesn’t have anything dramatic to announce yet and they aren’t willing to have live within the bounds of a forcing function like the PDC (PDC forces teams to get their acts together and finish off stuff enough to at least get some good demos together).

The last few PDCs haven’t exactly been huge successes, though. Hailstorm was announced at one and later was killed. Longhorn was announced at another and later was delayed and many things that were shown off were later killed too.

Now that Google, Amazon, Apple, are shipping platforms that are more and more interesting to Microsoft’s developer community Microsoft has to play a different game. One where they can’t keep showing off stuff that never ships. The stakes are going up in the Internet game and Microsoft doesn’t seem to have a good answer to what’s coming next.

Some other things I’m hearing from the Windows team? That they are still planning out the next version of Windows. So, I don’t expect to see a beta until 2008 (probably second half of the year, if we see one at all) and I don’t expect to see a major new version of Windows to ship until 2009.

Anyway, this is sad cause I was hoping to see Microsoft make an all out push for developers this year.

What do you think it all means? Am I reading too much in between the lines?

Comments

  1. Scoble said “This is a bummer, because .NET is a lot easier to write than C++ ”

    What you don’t seem to understand is that writing the code is just a slice of the development lifecycle. The total cost comprises a lot of other pieces.

    And I thought you linked to guys like Jeff Atwood or Scott Hanselman for a reason… /sarcasm

  2. Scoble said “This is a bummer, because .NET is a lot easier to write than C++ ”

    What you don’t seem to understand is that writing the code is just a slice of the development lifecycle. The total cost comprises a lot of other pieces.

    And I thought you linked to guys like Jeff Atwood or Scott Hanselman for a reason… /sarcasm

  3. Seriously though is this really a surprise?

    Mix07 was filled with so much cool new stuff it might as well have been a PDC?

    I was wondering what they had left over to unveil at the real PDC…

  4. Seriously though is this really a surprise?

    Mix07 was filled with so much cool new stuff it might as well have been a PDC?

    I was wondering what they had left over to unveil at the real PDC…

  5. Stephane: true. But there aren’t many projects that are starting today based on C++. Why not? Because other methodologies are faster/cheaper etc. Not to mention that most developers coming out of school are trained on Java and haven’t built a huge amount of expertise on C++.

    Jeff and Scott are both .NET types. That is if they are still writing Microsoft-centric code (I think Jeff wrote that he’s moving to LAMP and picking up Ruby on Rails).

    I had many developers inside Microsoft tell me that they wish they could use .NET instead of compiled C++. Why? It’s more fun to write .NET code and it’s faster. It’s also harder to write bugs due to the memory management.

    But, yeah, I know it is only part of the decision process. Thanks for pointing out to everyone that there’s more to the story than the writing code part.

    I assume my readers are smart and don’t need to have everything spelled out in inane detail like that.

  6. Stephane: true. But there aren’t many projects that are starting today based on C++. Why not? Because other methodologies are faster/cheaper etc. Not to mention that most developers coming out of school are trained on Java and haven’t built a huge amount of expertise on C++.

    Jeff and Scott are both .NET types. That is if they are still writing Microsoft-centric code (I think Jeff wrote that he’s moving to LAMP and picking up Ruby on Rails).

    I had many developers inside Microsoft tell me that they wish they could use .NET instead of compiled C++. Why? It’s more fun to write .NET code and it’s faster. It’s also harder to write bugs due to the memory management.

    But, yeah, I know it is only part of the decision process. Thanks for pointing out to everyone that there’s more to the story than the writing code part.

    I assume my readers are smart and don’t need to have everything spelled out in inane detail like that.

  7. Scoble said “After all, if Microsoft is unwilling to use it to develop Windows or Office, why should the rest of us base our life on it? Easy, it’s a lot more productive for the rest of us to write code in .NET”

    Again, no, the reason why Microsoft wants everyone to use .NET is because it’s tied to Windows. A .NET developer is a developer who’s using Windows. In and of itself, he’s a marketer for Windows. Any product shipped under that umbrella is an incentive to buy Windows licenses as opposed to Mac/Linux licenses.

    Please don’t get me started with the Linux .NET Mono thing. It’s from Novell, the company that will go bankrupt within two years for screwing their own customers…à la Microsoft.

  8. Scoble said “After all, if Microsoft is unwilling to use it to develop Windows or Office, why should the rest of us base our life on it? Easy, it’s a lot more productive for the rest of us to write code in .NET”

    Again, no, the reason why Microsoft wants everyone to use .NET is because it’s tied to Windows. A .NET developer is a developer who’s using Windows. In and of itself, he’s a marketer for Windows. Any product shipped under that umbrella is an incentive to buy Windows licenses as opposed to Mac/Linux licenses.

    Please don’t get me started with the Linux .NET Mono thing. It’s from Novell, the company that will go bankrupt within two years for screwing their own customers…à la Microsoft.

  9. Alex? So much cool stuff? Silverlight indeed was cool but was a repurposing of .NET and wasn’t that major. That’s not like a new version of Windows or Office. Maybe I’m reading it wrong, though.

  10. Alex? So much cool stuff? Silverlight indeed was cool but was a repurposing of .NET and wasn’t that major. That’s not like a new version of Windows or Office. Maybe I’m reading it wrong, though.

  11. Scoble said “Not to mention that most developers coming out of school are trained on Java and haven’t built a huge amount of expertise on C++.”

    You are probably right, but I think most are trained in academic Java (i.e. algorithms, infrastructure, …), not in the Java that is meant to make them become Java sweatshop workers.

  12. Scoble said “Not to mention that most developers coming out of school are trained on Java and haven’t built a huge amount of expertise on C++.”

    You are probably right, but I think most are trained in academic Java (i.e. algorithms, infrastructure, …), not in the Java that is meant to make them become Java sweatshop workers.

  13. Scoble said “subset of .NET is coming to Macintosh. Check out Silverlight”

    Scoble, you are screwing it again.

    Let me say for you : Microsoft Mac silverlight runtime is a free publicity. In a time where they are, arguably, being criticized for promoting stuff that only works with their own products/APIs, they have to come up with something that changes that perception. Repeat, PERCEPTION. What they want is to sell Windows licenses period.

    It’s foolish to think the Mac runtime is meant to show how good the Microsoft developers are, when at the same time they have every reason to build on their own investments, for instance DirectX (hardware acceleration in particular). This won’t ship for the Mac, right?

  14. Scoble said “subset of .NET is coming to Macintosh. Check out Silverlight”

    Scoble, you are screwing it again.

    Let me say for you : Microsoft Mac silverlight runtime is a free publicity. In a time where they are, arguably, being criticized for promoting stuff that only works with their own products/APIs, they have to come up with something that changes that perception. Repeat, PERCEPTION. What they want is to sell Windows licenses period.

    It’s foolish to think the Mac runtime is meant to show how good the Microsoft developers are, when at the same time they have every reason to build on their own investments, for instance DirectX (hardware acceleration in particular). This won’t ship for the Mac, right?

  15. Stephane: which is why I said “subset.” Of course the Mac will always be treated as a second class citizen. Silverlight was about beating Adobe’s Flex/Flash. I don’t remember DirectX being built into Flash either.

  16. Stephane: which is why I said “subset.” Of course the Mac will always be treated as a second class citizen. Silverlight was about beating Adobe’s Flex/Flash. I don’t remember DirectX being built into Flash either.

  17. For Microsoft Mac silverlight proposition to be taken seriously, they have to build something that looks like a Mac component, ie with the theming, look and feel, and all that stuff a Mac user expects.

    That’s why Microsoft insistence on reinventing the wheel inside the browser is not right. And we’ve seen that movie before with ActiveX controls.

  18. For Microsoft Mac silverlight proposition to be taken seriously, they have to build something that looks like a Mac component, ie with the theming, look and feel, and all that stuff a Mac user expects.

    That’s why Microsoft insistence on reinventing the wheel inside the browser is not right. And we’ve seen that movie before with ActiveX controls.

  19. According to the post on Microsoft’s PDC site it says the event is being re-scheduled (aka postponed) and not cancelled. No where do I see Microsoft saying the event is cancelled.

    Perhaps they will hold it sometime in early Winter? Maybe Spring 2008?

    For an event where many folks were looking forward to, I highly doubt it has been cancelled.

  20. According to the post on Microsoft’s PDC site it says the event is being re-scheduled (aka postponed) and not cancelled. No where do I see Microsoft saying the event is cancelled.

    Perhaps they will hold it sometime in early Winter? Maybe Spring 2008?

    For an event where many folks were looking forward to, I highly doubt it has been cancelled.

  21. Vista has some seriously obvious bugs all over the shell, and I can’t help but think that had they used WPF (with it’s excellent databinding/refresh) to write a clean version of the shell (and I’m really not interested in hearing MS’ excuses about code investment…the shit’s broken, and it’s got too many layers of functionality to clean it up properly; it was bad in XP, but in Vista, it’s an embarrassment).
    I can see them not wanting to have managed drivers or lower-level stuff, but the shell?!?! Hell, I’ve seen a friggin’ AJAX approximation of the core functionality of Aero, so it can’t be a huge stretch. But regardless, I’ve seen this in every app I’ve worked on: every rev, more features get added on & bugs get fixed, but refactoring is too risky/too costly/too something, and the cohesiveness of what the app should be gets lost as hacks get thrown in, or features that should use existing code get reimplemented separately.

  22. Vista has some seriously obvious bugs all over the shell, and I can’t help but think that had they used WPF (with it’s excellent databinding/refresh) to write a clean version of the shell (and I’m really not interested in hearing MS’ excuses about code investment…the shit’s broken, and it’s got too many layers of functionality to clean it up properly; it was bad in XP, but in Vista, it’s an embarrassment).
    I can see them not wanting to have managed drivers or lower-level stuff, but the shell?!?! Hell, I’ve seen a friggin’ AJAX approximation of the core functionality of Aero, so it can’t be a huge stretch. But regardless, I’ve seen this in every app I’ve worked on: every rev, more features get added on & bugs get fixed, but refactoring is too risky/too costly/too something, and the cohesiveness of what the app should be gets lost as hacks get thrown in, or features that should use existing code get reimplemented separately.

  23. MIX this year was full of exciting announcements for developers. A PDC before the end of the year could not add much.

    Also, many teams in the Developer Division have been shipping early CTPs of their products for some time, but they have had a few bad steps (especially with WinFS) and now they cannot even announce something is going to be delayed without putting all developers in panic.

    So I guess they are learning to be more careful. Personally, I like it more if they blow my expectations when they ship :)

    Regarding .NET code in Windows, it is an interesting topic. I think you will see much more of it in Windows Server 2008. By the way, I wonder how much of Solaris is written in Java these days.

  24. MIX this year was full of exciting announcements for developers. A PDC before the end of the year could not add much.

    Also, many teams in the Developer Division have been shipping early CTPs of their products for some time, but they have had a few bad steps (especially with WinFS) and now they cannot even announce something is going to be delayed without putting all developers in panic.

    So I guess they are learning to be more careful. Personally, I like it more if they blow my expectations when they ship :)

    Regarding .NET code in Windows, it is an interesting topic. I think you will see much more of it in Windows Server 2008. By the way, I wonder how much of Solaris is written in Java these days.

  25. As I understand it, the problem with using .net internally in Vista is that processes can’t use more than one version of the .net runtime. So that pretty much eliminates it from usage in any DLL that developers would link to, because otherwise we’d be locked into whatever version of the runtime they developed against.

    I suspect complete, standalone apps could be written in .net with no problems, but there really isn’t much of that in the OS.

    Note: I was a c++ developer for years, but have been developing in c# exclusively for about the last 4. I wouldn’t have a problem going back to c++. But I’d hate to have to work with new guys coming up in c++ who haven’t had the practices beat into them that avoid c++ problems like memory leaks. c# avoids some (but not all) of that.

  26. As I understand it, the problem with using .net internally in Vista is that processes can’t use more than one version of the .net runtime. So that pretty much eliminates it from usage in any DLL that developers would link to, because otherwise we’d be locked into whatever version of the runtime they developed against.

    I suspect complete, standalone apps could be written in .net with no problems, but there really isn’t much of that in the OS.

    Note: I was a c++ developer for years, but have been developing in c# exclusively for about the last 4. I wouldn’t have a problem going back to c++. But I’d hate to have to work with new guys coming up in c++ who haven’t had the practices beat into them that avoid c++ problems like memory leaks. c# avoids some (but not all) of that.

  27. “the problem with using .net internally in Vista is that processes can’t use more than one version of the .net runtime.”

    It’s indeed one of the problems, even though a fix for that is to proxy calls to outside EXEs. Doing so means that you are back to good ol’ COM infrastructure…

    What Microsoft faces with this (remember: it’s just one problem), is also faced by anybody out there. For instance, if you are deployed an Excel addin written in .NET, you are at the mercy of everyone else’s Excel addin which are loaded before you. It’s a non-starter except for trivial addins that are version independent.

    “But I’d hate to have to work with new guys coming up in c++ who haven’t had the practices beat into them that avoid c++ problems like memory leaks. c# avoids some (but not all) of that.”

    C++ memory leaks? Sure, if you are not disciplined, it will burn you. But frankly, if you are building on a proper infrastructure, all your allocations are properly wrapped, and it does not leak.
    In .NET, I would not say this thing does not leak. First because there is the perception of leak, just do a loop that creates new Strings, and because if your code (or a library you depend on) uses any COM/PInvoke transition, then you are back to square one. For instance, System.Drawing is a wrapper of GDI+. GDI/GDI+ is severely limited by what’s known as the “desktop heap”. All the wrong conclusions follow, despite that you are using .NET

  28. “the problem with using .net internally in Vista is that processes can’t use more than one version of the .net runtime.”

    It’s indeed one of the problems, even though a fix for that is to proxy calls to outside EXEs. Doing so means that you are back to good ol’ COM infrastructure…

    What Microsoft faces with this (remember: it’s just one problem), is also faced by anybody out there. For instance, if you are deployed an Excel addin written in .NET, you are at the mercy of everyone else’s Excel addin which are loaded before you. It’s a non-starter except for trivial addins that are version independent.

    “But I’d hate to have to work with new guys coming up in c++ who haven’t had the practices beat into them that avoid c++ problems like memory leaks. c# avoids some (but not all) of that.”

    C++ memory leaks? Sure, if you are not disciplined, it will burn you. But frankly, if you are building on a proper infrastructure, all your allocations are properly wrapped, and it does not leak.
    In .NET, I would not say this thing does not leak. First because there is the perception of leak, just do a loop that creates new Strings, and because if your code (or a library you depend on) uses any COM/PInvoke transition, then you are back to square one. For instance, System.Drawing is a wrapper of GDI+. GDI/GDI+ is severely limited by what’s known as the “desktop heap”. All the wrong conclusions follow, despite that you are using .NET

  29. “But almost no Vista code was written in .NET (if any, actually). Microsoft tries to keep this secret because they know it gives a black eye to .NET”

    “The original plan for Longhorn was to build lots of components on top of the next version of the .Net Framework,” according to one of our developer sources, who requested anonymity. “But given how late (.Net Framework 2.0) is, and how new it would be (Microsoft Chairman) Bill Gates realized it would be foolish to build important pieces of Longhorn on top of .Net.”
    (http://www.microsoft-watch.com/content/developer/the_dirty_little_secret_about_longhorn.html)

    – I believe a lot of Vistas Media Center was written in managed code.

    The following blog entries outline some of the other MS products using managed code.

    http://blogs.msdn.com/danielfe/archive/2005/12/16/504847.aspx
    http://blogs.msdn.com/danielfe/archive/2004/11/02/251254.aspx

    I definitely think a managed kernel wont be in Windows for a while, but services within windows itself could take advantage of managed code.

  30. “But almost no Vista code was written in .NET (if any, actually). Microsoft tries to keep this secret because they know it gives a black eye to .NET”

    “The original plan for Longhorn was to build lots of components on top of the next version of the .Net Framework,” according to one of our developer sources, who requested anonymity. “But given how late (.Net Framework 2.0) is, and how new it would be (Microsoft Chairman) Bill Gates realized it would be foolish to build important pieces of Longhorn on top of .Net.”
    (http://www.microsoft-watch.com/content/developer/the_dirty_little_secret_about_longhorn.html)

    – I believe a lot of Vistas Media Center was written in managed code.

    The following blog entries outline some of the other MS products using managed code.

    http://blogs.msdn.com/danielfe/archive/2005/12/16/504847.aspx
    http://blogs.msdn.com/danielfe/archive/2004/11/02/251254.aspx

    I definitely think a managed kernel wont be in Windows for a while, but services within windows itself could take advantage of managed code.

  31. Scoble said “But almost no Vista code was written in .NET (if any, actually). ”

    I thought Media Center and some of the applets were written in C#?

    I’m not sure what the issue is with performance, I use Expression Blend all the time and it seems just fine to me (and it is all WPF). Sure, you wouldn’t want to use C# for core bits of the OS but for apps it would seem OK.

    Before they worry about competing with OSX in terms of performance they need to compete in terms of security and usability.

    You are coming to a sad realization. Cancel or allow?

  32. Scoble said “But almost no Vista code was written in .NET (if any, actually). ”

    I thought Media Center and some of the applets were written in C#?

    I’m not sure what the issue is with performance, I use Expression Blend all the time and it seems just fine to me (and it is all WPF). Sure, you wouldn’t want to use C# for core bits of the OS but for apps it would seem OK.

    Before they worry about competing with OSX in terms of performance they need to compete in terms of security and usability.

    You are coming to a sad realization. Cancel or allow?

  33. Actually, Stephane…it’s idiotic to state that you’ll avoid memory leaks in C++ if you “stay disciplined”. The same thing obviously applies to C#. The same applies to .Net. You’ll avoid memory leaks if you also “stay disciplined”. Forget the marketing literature. Personally, I can’t think of any our current .Net projects…we have at least 15 or so large scale systems out in the wild today…that have memory leak issues. I shudder to think how long it would have taken to reach feature parity on these projects had we used C++.

    Oh and Robert, I think you’ll find that most .Net devs don’t actually care all that much that there is no “OMG!! another new dev platform!!!!1″, if anything, we’ve been barraged with them for the past few years and it’s hard to keep up.

    Non story if you ask me.

  34. Actually, Stephane…it’s idiotic to state that you’ll avoid memory leaks in C++ if you “stay disciplined”. The same thing obviously applies to C#. The same applies to .Net. You’ll avoid memory leaks if you also “stay disciplined”. Forget the marketing literature. Personally, I can’t think of any our current .Net projects…we have at least 15 or so large scale systems out in the wild today…that have memory leak issues. I shudder to think how long it would have taken to reach feature parity on these projects had we used C++.

    Oh and Robert, I think you’ll find that most .Net devs don’t actually care all that much that there is no “OMG!! another new dev platform!!!!1″, if anything, we’ve been barraged with them for the past few years and it’s hard to keep up.

    Non story if you ask me.

  35. Contrast this to Google who is holding a huge developer day next week (it sold out, so I won’t even bother linking to it). Or Facebook, who held a big developer-centric shindig today.

    You are really comparing the incomparables here. Like a few people pointed out in the comments here, Mix ’07 was a huge event and there’s a lot going on with respect to DLR. There’s nothing half as exciting IMHO, from Google land.

    As for the Facebook announcement, Microsoft also did a demo there talking about Popfly integration in Facebook.

    The PDC being postponed does not indicate a slack in what Microsoft does for developers. You are just reading too much between the lines.

  36. Contrast this to Google who is holding a huge developer day next week (it sold out, so I won’t even bother linking to it). Or Facebook, who held a big developer-centric shindig today.

    You are really comparing the incomparables here. Like a few people pointed out in the comments here, Mix ’07 was a huge event and there’s a lot going on with respect to DLR. There’s nothing half as exciting IMHO, from Google land.

    As for the Facebook announcement, Microsoft also did a demo there talking about Popfly integration in Facebook.

    The PDC being postponed does not indicate a slack in what Microsoft does for developers. You are just reading too much between the lines.

  37. Of course the Mac will always be treated as a second class citizen. Silverlight was about beating Adobe’s Flex/Flash. I don’t remember DirectX being built into Flash either.

    That’s because you can’t run Direct X on anything but Windows. For Direct X to be in Flash would be quite stupid.

  38. Of course the Mac will always be treated as a second class citizen. Silverlight was about beating Adobe’s Flex/Flash. I don’t remember DirectX being built into Flash either.

    That’s because you can’t run Direct X on anything but Windows. For Direct X to be in Flash would be quite stupid.

  39. “Alex? So much cool stuff? Silverlight indeed was cool but was a repurposing of .NET and wasn’t that major.”

    Ah but I think you are missing the importance of things like Astoria… I have pretty good idea of the long term direction of Astoria and it is pretty much going to take us into the realm of Intentional Programming. I have a whole series of posts on this stretching back over a year now. This was the first: http://www.base4.net/blog.aspx?ID=36

    See the value in REST is that it forces you to create intentional systems. And there is no better way to integrate systems… and some people at Microsoft get that now.

    More importantly intentional systems plus meta-data about those systems will make it possible to create smart interfaces that allow for end-users not programmers to program systems (albeit by simply expressing their intent), the possibilities for the end-user experience and giving power to users are enormous…

    but then maybe I am getting ahead of myself!

  40. “Alex? So much cool stuff? Silverlight indeed was cool but was a repurposing of .NET and wasn’t that major.”

    Ah but I think you are missing the importance of things like Astoria… I have pretty good idea of the long term direction of Astoria and it is pretty much going to take us into the realm of Intentional Programming. I have a whole series of posts on this stretching back over a year now. This was the first: http://www.base4.net/blog.aspx?ID=36

    See the value in REST is that it forces you to create intentional systems. And there is no better way to integrate systems… and some people at Microsoft get that now.

    More importantly intentional systems plus meta-data about those systems will make it possible to create smart interfaces that allow for end-users not programmers to program systems (albeit by simply expressing their intent), the possibilities for the end-user experience and giving power to users are enormous…

    but then maybe I am getting ahead of myself!

  41. Robert your sooo predictable,I told you before,get out of the Valley, your becoming a cheerleader for Google and any other company in the area, just as bad as when your worked for MSFT.

    Move to Texas or something,no then you would be preaching to the masses about Dell and the Dallas Cowboys.

  42. Robert your sooo predictable,I told you before,get out of the Valley, your becoming a cheerleader for Google and any other company in the area, just as bad as when your worked for MSFT.

    Move to Texas or something,no then you would be preaching to the masses about Dell and the Dallas Cowboys.

  43. “The person who told me this (who works at Microsoft) told me .NET still takes too long to startup and load into memory and because Windows is now being compared to OSX they can’t afford to ship components that would slow down Windows.”

    Robert Scoble

    Are you kidding me? how could that hurt,OSX may be a nice OS, but fast isn’t something that comes to mind.

  44. “The person who told me this (who works at Microsoft) told me .NET still takes too long to startup and load into memory and because Windows is now being compared to OSX they can’t afford to ship components that would slow down Windows.”

    Robert Scoble

    Are you kidding me? how could that hurt,OSX may be a nice OS, but fast isn’t something that comes to mind.

  45. >>The person who told me this (who works at Microsoft) told me .NET still takes too long to startup and load into memory

    This is one of the big reasons why many of us are still supporting asp and com based web applications.

  46. >>The person who told me this (who works at Microsoft) told me .NET still takes too long to startup and load into memory

    This is one of the big reasons why many of us are still supporting asp and com based web applications.

  47. Are you kidding me? An OS built with .net? Why would you think that an app framework could be used to build an OS? Man, take the time to understand what you’re saying here. An OS is not in any way, shape or form like some garage-based web 2.0 project built by a 4th grader. An OS is serious stuff and has different and special requirements that plain old apps don’t.

  48. Are you kidding me? An OS built with .net? Why would you think that an app framework could be used to build an OS? Man, take the time to understand what you’re saying here. An OS is not in any way, shape or form like some garage-based web 2.0 project built by a 4th grader. An OS is serious stuff and has different and special requirements that plain old apps don’t.

  49. Thanks for sharing this info with us Scoble.

    Its funny both Java and .NET are byte code based languages and Sun on onther hand is trying they can to push JRE as platform but doesnt seem to have lot of success except some success in mobile phone market, which in my opinion is already crowded. I do not have any benchmark on how well Java performs but given how far .NET has come (specially with 3.0 and 3.5) I am still betting my horses on .NET

    I recently talked to one of the CLR team members and they certainy are aware of this issue and my impresssion was that they are close to solving the loading issue. We have to wait and see how things folds in next 6-12 months.

  50. Thanks for sharing this info with us Scoble.

    Its funny both Java and .NET are byte code based languages and Sun on onther hand is trying they can to push JRE as platform but doesnt seem to have lot of success except some success in mobile phone market, which in my opinion is already crowded. I do not have any benchmark on how well Java performs but given how far .NET has come (specially with 3.0 and 3.5) I am still betting my horses on .NET

    I recently talked to one of the CLR team members and they certainy are aware of this issue and my impresssion was that they are close to solving the loading issue. We have to wait and see how things folds in next 6-12 months.

  51. We have dual core processors on our laptops today, in a year we’ll have quad core and soon after eight core.

    Multi-core, multi-threaded managed code is our developer platform destiny.

  52. We have dual core processors on our laptops today, in a year we’ll have quad core and soon after eight core.

    Multi-core, multi-threaded managed code is our developer platform destiny.

  53. What I meant about C++ is that, over years of doing c++ little techniques to manage the complexity become second nature. Things like ‘oh, I just created a heap object here, before I write the code that uses it I’m going to jump to the disposal spot and write the delete call’. Or ‘hmm, there is no one place to put the delete call. That’s a bad idea, let me redesign this.’ There are dozens, if not hundreds of things like this that any seasoned decent c++ guy will do automatically, because he’s had to spend time fixing the code that didn’t do them.

    Guys who’ve never spent time in the trenches in a non-GC language don’t have this. So if they do end up having to work in c++, even if they’re now experienced developers I’m going to have to treat them like entry-level guys.

  54. What I meant about C++ is that, over years of doing c++ little techniques to manage the complexity become second nature. Things like ‘oh, I just created a heap object here, before I write the code that uses it I’m going to jump to the disposal spot and write the delete call’. Or ‘hmm, there is no one place to put the delete call. That’s a bad idea, let me redesign this.’ There are dozens, if not hundreds of things like this that any seasoned decent c++ guy will do automatically, because he’s had to spend time fixing the code that didn’t do them.

    Guys who’ve never spent time in the trenches in a non-GC language don’t have this. So if they do end up having to work in c++, even if they’re now experienced developers I’m going to have to treat them like entry-level guys.

  55. You are off on this one.

    PDC is about *developers*. There is a huge new release of Visual Studio coming in the fall (“Orcas”), and a huge new release of SQL server (“katmai”).

    These things are the bread and butter of PDC. This is the PERFECT time to have a PDC. There is absolutely no reason for MS to cancel it unless they are really floundering.

  56. You are off on this one.

    PDC is about *developers*. There is a huge new release of Visual Studio coming in the fall (“Orcas”), and a huge new release of SQL server (“katmai”).

    These things are the bread and butter of PDC. This is the PERFECT time to have a PDC. There is absolutely no reason for MS to cancel it unless they are really floundering.

  57. That shows you how much confidence they have in silverflash!

    They’re going down, down, down…

  58. That shows you how much confidence they have in silverflash!

    They’re going down, down, down…

  59. When Steve Sinofsky ran the Office team, it was difficult if not impossible to get information out them. They ran a tight ship with little leaks. Why are we surprised at not hearing much about the next Windows now that he runs that group?

  60. When Steve Sinofsky ran the Office team, it was difficult if not impossible to get information out them. They ran a tight ship with little leaks. Why are we surprised at not hearing much about the next Windows now that he runs that group?

  61. The cancelation has more to do with Ray Ozzie and Steven Sinofsky’s desire for secrecy more than anything
    else.
    The days of MSFT giving the longview of their coming products is over.

  62. The cancelation has more to do with Ray Ozzie and Steven Sinofsky’s desire for secrecy more than anything
    else.
    The days of MSFT giving the longview of their coming products is over.

  63. “This is a bummer, because .NET is a lot easier to write than C++ ”

    You obviously don’t know this from experience. Aren’t you basically repeating things that others have told you? You DO understand the differences between Ç++ and .Net, right? The fact that one is a programming language and one is a development platform? Perhaps you meant to compare the differences between C++ and C#?

  64. “This is a bummer, because .NET is a lot easier to write than C++ ”

    You obviously don’t know this from experience. Aren’t you basically repeating things that others have told you? You DO understand the differences between Ç++ and .Net, right? The fact that one is a programming language and one is a development platform? Perhaps you meant to compare the differences between C++ and C#?

  65. Regarding the C++ practices of memory cleanup carrying over to C#, sure if you’ve never had to craft your memory allocs to use more stack allocs vs. heap allocs for auto-cleanup or pay attention to delete’ing your new’ed vars, you can still code leakless C# apps. However, the aforementioned practices are still valuable in .Net as a whole. Think about the use of the “using” statement in conjunction with IDisposable for the former scenario. And for memory leaks, you may not get those if you aren’t accustomed to cleaning up your own allocs, but you can still get leaks of other, non-GC’ed resources, such as database handles and COM object references. You can effectively orphan these things even with the GC’er.

  66. Regarding the C++ practices of memory cleanup carrying over to C#, sure if you’ve never had to craft your memory allocs to use more stack allocs vs. heap allocs for auto-cleanup or pay attention to delete’ing your new’ed vars, you can still code leakless C# apps. However, the aforementioned practices are still valuable in .Net as a whole. Think about the use of the “using” statement in conjunction with IDisposable for the former scenario. And for memory leaks, you may not get those if you aren’t accustomed to cleaning up your own allocs, but you can still get leaks of other, non-GC’ed resources, such as database handles and COM object references. You can effectively orphan these things even with the GC’er.

  67. Brandon: “According to the post on Microsoft’s PDC site it says the event is being re-scheduled (aka postponed) and not cancelled. ”

    Uh, yeah. Ranks right up there with “I still want to be friends” and “the check is in your mouth.” If you say “cancelled” I bet you have to give back all the money from the sponsors and exhibitors. I bet you have to actually admit that you couldn’t put on a meaningful show and that customers would rip you a new one. I bet that you’d have a great platform to promote Live! and for some reason just don’t want to do it.

    Lines. Read between. Proof is left for the student as an exercise.

  68. Brandon: “According to the post on Microsoft’s PDC site it says the event is being re-scheduled (aka postponed) and not cancelled. ”

    Uh, yeah. Ranks right up there with “I still want to be friends” and “the check is in your mouth.” If you say “cancelled” I bet you have to give back all the money from the sponsors and exhibitors. I bet you have to actually admit that you couldn’t put on a meaningful show and that customers would rip you a new one. I bet that you’d have a great platform to promote Live! and for some reason just don’t want to do it.

    Lines. Read between. Proof is left for the student as an exercise.

  69. LayZ: I used to work as an editor at Visual Basic Programmer’s Journal so I know the difference between C++ that compiles to native code and VB or C# that runs on the .NET compiler/framework/ecosystem or whatever you want to call it. But, you’re right. Most of this post comes from what other people who know a lot more than I do tell me.

  70. LayZ: I used to work as an editor at Visual Basic Programmer’s Journal so I know the difference between C++ that compiles to native code and VB or C# that runs on the .NET compiler/framework/ecosystem or whatever you want to call it. But, you’re right. Most of this post comes from what other people who know a lot more than I do tell me.

  71. “This is a bummer, because .NET is a lot easier to write than C++”

    Though this statement is incorrect in a strict sense(you can’t write IN .NET) it is also very true in another sense.

    irrespective of the language you use(VB,C#, Managed C++ …) programming with the .NET FW and targeting the CLR is easier than writing native code using C++.

    The general statement “Writing apps using .NET is easier than C++” is still very valid.

  72. “This is a bummer, because .NET is a lot easier to write than C++”

    Though this statement is incorrect in a strict sense(you can’t write IN .NET) it is also very true in another sense.

    irrespective of the language you use(VB,C#, Managed C++ …) programming with the .NET FW and targeting the CLR is easier than writing native code using C++.

    The general statement “Writing apps using .NET is easier than C++” is still very valid.

  73. Geeze.
    They JUST had a very successful MIX conference, which was a *real* DC. Scoble wants “DCs” like Apple’s WWDCs, which, despite the “DC” in the name, are aimed at consumers rather than devs. (Just compare the MIX07 keynote to a Jobs WWDC keynote, and see which is a real “DC”.) Scoble, you say you want announcements of new Windows and Office? New versions JUST now came out.

    And the notion that Microsoft is doing nothing for devs this year is more foolishness. Ever heard of Orcas? Ever heard of LINQ? Ever heard of C# 3.0 and 3.5? You should read Wes Deyer’s MSDN blog for some of the cool C# features that are coming this year (whcih will put Java more behind than it already is, feature wise).

    Lastly, your “informants” are pulling your leg. OSX isn’t known for its speed, so why would anyone be worried about speed comparisons with OSX?

    And this “.NET takes too long to load” is bull. Once the first app loads it, all other apps that use it load very quickly. The entire Expression sweet is WPF for crying out loud. So Microsoft is using .NET for new apps (expecting them to rewrite Office in .NET is idiotic).

    I don’t know where Scoble is getting his info; either from people taking advantage of his gullibility by pulling his leg, or from the malcontents at minimsft.

  74. Geeze.
    They JUST had a very successful MIX conference, which was a *real* DC. Scoble wants “DCs” like Apple’s WWDCs, which, despite the “DC” in the name, are aimed at consumers rather than devs. (Just compare the MIX07 keynote to a Jobs WWDC keynote, and see which is a real “DC”.) Scoble, you say you want announcements of new Windows and Office? New versions JUST now came out.

    And the notion that Microsoft is doing nothing for devs this year is more foolishness. Ever heard of Orcas? Ever heard of LINQ? Ever heard of C# 3.0 and 3.5? You should read Wes Deyer’s MSDN blog for some of the cool C# features that are coming this year (whcih will put Java more behind than it already is, feature wise).

    Lastly, your “informants” are pulling your leg. OSX isn’t known for its speed, so why would anyone be worried about speed comparisons with OSX?

    And this “.NET takes too long to load” is bull. Once the first app loads it, all other apps that use it load very quickly. The entire Expression sweet is WPF for crying out loud. So Microsoft is using .NET for new apps (expecting them to rewrite Office in .NET is idiotic).

    I don’t know where Scoble is getting his info; either from people taking advantage of his gullibility by pulling his leg, or from the malcontents at minimsft.

  75. You’ve got to be kidding if you think the keynote with Mike Arrington and Ray Ozzie at Mix07 was any more developer focused than what Steve Jobs does at the WWDC.

    I’ve heard the .NET startup time problem from a range of developers across Microsoft’s Windows team. It isn’t a big deal to wait for Silverlight to startup, but it IS a huge deal to delay Windows starting up by a few seconds.

    Ask yourself: why isn’t any code in Windows written in .NET? Not even the games were written in .NET. That’s a fact and is one that I hear probably won’t change, even in the next version of Windows.

  76. OSX starts up fast, especially off of sleep mode. I have a Mac and it starts up faster than my Vista machines (and is more reliable on startup too).

  77. You’ve got to be kidding if you think the keynote with Mike Arrington and Ray Ozzie at Mix07 was any more developer focused than what Steve Jobs does at the WWDC.

    I’ve heard the .NET startup time problem from a range of developers across Microsoft’s Windows team. It isn’t a big deal to wait for Silverlight to startup, but it IS a huge deal to delay Windows starting up by a few seconds.

    Ask yourself: why isn’t any code in Windows written in .NET? Not even the games were written in .NET. That’s a fact and is one that I hear probably won’t change, even in the next version of Windows.

  78. OSX starts up fast, especially off of sleep mode. I have a Mac and it starts up faster than my Vista machines (and is more reliable on startup too).

  79. And, David, maybe you should do some research into what I did at Microsoft. Go and look at this video of Anders Hejlsberg. It was where he announced LINQ to the world. I was behind the camera. http://channel9.msdn.com/ShowPost.aspx?PostID=115010
    Oh, that was September 2005. So, where’s the “new” stuff? That’s what the PDC is for. When Microsoft decides not to hold a PDC it’s an admission that they don’t have enough “new” stuff to talk about.

  80. And, David, maybe you should do some research into what I did at Microsoft. Go and look at this video of Anders Hejlsberg. It was where he announced LINQ to the world. I was behind the camera. http://channel9.msdn.com/ShowPost.aspx?PostID=115010
    Oh, that was September 2005. So, where’s the “new” stuff? That’s what the PDC is for. When Microsoft decides not to hold a PDC it’s an admission that they don’t have enough “new” stuff to talk about.

  81. Robert, quite with the “whyt isn’t Windows written in .Net”.

    Let me ask you a question, what cool language that Google/Technorati/Flickr/Twitter (or whoever is “hip and happening” in your world today) uses should Windows components be written in? Javascript? Ruby? Python? Java? What are you trying to say exactly? What language/platform should they use in your highly esteemed opinion?

  82. Robert, quite with the “whyt isn’t Windows written in .Net”.

    Let me ask you a question, what cool language that Google/Technorati/Flickr/Twitter (or whoever is “hip and happening” in your world today) uses should Windows components be written in? Javascript? Ruby? Python? Java? What are you trying to say exactly? What language/platform should they use in your highly esteemed opinion?

  83. Using an app framework for an OS isn’t the right tool for the job. There are .Net apps coming from MS,and there are apps that have a good deal of legacy code to work with that make more sense not using .Net – idealism shouldn’t run project management.

    We are far from the days of MFC where Microsoft wrote only Notepad in their own framework. It’s also wrong to say Microsoft is only out to sell Vista – that’s no different than saying Google is only a search engine.

    DevConnections and other conferences are still coming, and with MIX just over (and SilverLight running a .Net CLR on Win and Mac is a big deal) it can simply be there isn’t a need for PDC this year. Microsoft is doing excellent by developers with Orcas, .Net 3.5, LINQ, and the Expression lineup. As a blogger, it’s cool to talk about what’s coming – but as a developer it only matters what is here. This stuff may be old news to some, but to developers it’s real now – and that’s a big deal.

  84. Using an app framework for an OS isn’t the right tool for the job. There are .Net apps coming from MS,and there are apps that have a good deal of legacy code to work with that make more sense not using .Net – idealism shouldn’t run project management.

    We are far from the days of MFC where Microsoft wrote only Notepad in their own framework. It’s also wrong to say Microsoft is only out to sell Vista – that’s no different than saying Google is only a search engine.

    DevConnections and other conferences are still coming, and with MIX just over (and SilverLight running a .Net CLR on Win and Mac is a big deal) it can simply be there isn’t a need for PDC this year. Microsoft is doing excellent by developers with Orcas, .Net 3.5, LINQ, and the Expression lineup. As a blogger, it’s cool to talk about what’s coming – but as a developer it only matters what is here. This stuff may be old news to some, but to developers it’s real now – and that’s a big deal.

  85. I well remember the 1996 PDC, where almost every keynote announcement (including Microsoft signing a licensing deal with Sun for Java) was greeted with wild applause. And there was Jobs on stage as a speaker, 100% uncomfortable and insincere, but … THERE.

    The “100 Days into the Internet” vibe of that PDC just made it an extraordinary event.

    Cut to the 2005 PDC, where virtually the whole of BillG’s and Allchin’s keynotes were received in stoney silence (Speakers has to resort to begging for applause: “Oh, so you’re not impressed by that, huh?”).

    No fun for anyone, CTPs and MSFT blogging had already pre-empted most of it. Old news, all of it … at $4K a pop.

    A PDC in support of bit of Orcas (already beta’d at MIX) and a bit of Katmai would fall even flatter. Forget about it.

    In fact the whole PDC thing may even have run its course in favour of more lightweight gigs like MIX.

    We’ll see.

  86. I well remember the 1996 PDC, where almost every keynote announcement (including Microsoft signing a licensing deal with Sun for Java) was greeted with wild applause. And there was Jobs on stage as a speaker, 100% uncomfortable and insincere, but … THERE.

    The “100 Days into the Internet” vibe of that PDC just made it an extraordinary event.

    Cut to the 2005 PDC, where virtually the whole of BillG’s and Allchin’s keynotes were received in stoney silence (Speakers has to resort to begging for applause: “Oh, so you’re not impressed by that, huh?”).

    No fun for anyone, CTPs and MSFT blogging had already pre-empted most of it. Old news, all of it … at $4K a pop.

    A PDC in support of bit of Orcas (already beta’d at MIX) and a bit of Katmai would fall even flatter. Forget about it.

    In fact the whole PDC thing may even have run its course in favour of more lightweight gigs like MIX.

    We’ll see.

  87. I’ve long been a Windows user (don’t think I’ll go so far as fan) and developer, creating both client and Windows applications. The purchase of a notebook with Vista was the impetus for buying my first Mac. The speed of the mac prompted the purchase of a macbook three weeks later.

    Why is any of this relevant? Microsoft should be treading lightly with any new announcements as Vista is causing many long-time Windows users to throw their hands up in disgust and high-tail it the Apple Store to buy a computer that boots in 30 seconds and doesn’t run the hard drive incessantly for thirty minutes for no apparent reason. Was Vista really ready for primetime or did they just need a fresh influx of cash that quarter?

    I think .NET wasn’t used in the OS because they were already battling speed issues and .NET was only going to contribute to the problem.

  88. I’ve long been a Windows user (don’t think I’ll go so far as fan) and developer, creating both client and Windows applications. The purchase of a notebook with Vista was the impetus for buying my first Mac. The speed of the mac prompted the purchase of a macbook three weeks later.

    Why is any of this relevant? Microsoft should be treading lightly with any new announcements as Vista is causing many long-time Windows users to throw their hands up in disgust and high-tail it the Apple Store to buy a computer that boots in 30 seconds and doesn’t run the hard drive incessantly for thirty minutes for no apparent reason. Was Vista really ready for primetime or did they just need a fresh influx of cash that quarter?

    I think .NET wasn’t used in the OS because they were already battling speed issues and .NET was only going to contribute to the problem.

  89. Are you kidding me? how could that hurt,OSX may be a nice OS, but fast isn’t something that comes to mind.

    Wake-up call: it’s 2007 now. We’re talking versus Vista, not XP.

    I’ve run OS X on an old G3 with 512 MB of RAM. It’s certainly not fast, but it’s actually usable (for basic stuff – e-mail, web browsing, Office, etc.). I dare you to try using Vista on a Pentium II/366 with 512 MB of RAM and an old ATI Rage card. Even just using IE or Office would be an incredibly painful and slow experience (ignoring the fact that you probably wouldn’t even be able to get it to install or boot).

  90. Are you kidding me? how could that hurt,OSX may be a nice OS, but fast isn’t something that comes to mind.

    Wake-up call: it’s 2007 now. We’re talking versus Vista, not XP.

    I’ve run OS X on an old G3 with 512 MB of RAM. It’s certainly not fast, but it’s actually usable (for basic stuff – e-mail, web browsing, Office, etc.). I dare you to try using Vista on a Pentium II/366 with 512 MB of RAM and an old ATI Rage card. Even just using IE or Office would be an incredibly painful and slow experience (ignoring the fact that you probably wouldn’t even be able to get it to install or boot).

  91. “Not even the games were written in .NET.”

    WTF?!?!?! What kind of excuse do they have for *this* one?

  92. “Not even the games were written in .NET.”

    WTF?!?!?! What kind of excuse do they have for *this* one?

  93. .net does start up slowly, go browse some .net 1.1 web apps that have been sitting idle (no activity) for more then 20 minutes. my old vb com with asp beats that stuff with no startup wait and no configuration changes to make it go faster. why is that?

  94. .net does start up slowly, go browse some .net 1.1 web apps that have been sitting idle (no activity) for more then 20 minutes. my old vb com with asp beats that stuff with no startup wait and no configuration changes to make it go faster. why is that?

  95. that they don’t have enough “new” stuff to talk about.

    Oh they have tons to TALK about, delivery is another matter, too early in the hype spin cycle and I don’t think they want a PDC 2003 repeat.

  96. that they don’t have enough “new” stuff to talk about.

    Oh they have tons to TALK about, delivery is another matter, too early in the hype spin cycle and I don’t think they want a PDC 2003 repeat.

  97. “You’ve got to be kidding if you think the keynote with Mike Arrington and Ray Ozzie at Mix07 was any more developer focused than what Steve Jobs does at the WWDC.”

    Scoble, have you even bothered to watch the Mix07 keynote, which was mainly mc’ed by Scott Gutthrie, not Ozzie or Arrington. During the keynote, someone demoed actually creating and debugging a solution from start to finish. When has a WWDC *ever* showed debugging as part of a keynote? WWDCs are about, “look at this new Leopard feature!” nonsense that devs really don’t care about.

    “I’ve heard the .NET startup time problem from a range of developers across Microsoft’s Windows team. It isn’t a big deal to wait for Silverlight to startup, but it IS a huge deal to delay Windows starting up by a few seconds.”

    This is absurd. My own .NET apps start within 5 seconds if .NET hasn’t been loaded yet, and start instantaneously if it has been. And this has zero to do with waking from sleep mode. From where do you obtain your tech knowledge?

    And yes, I do know that you did LINQ interviews at Channel9. That doesn’t mean you actually know anything about the tech. Most of your questions are of the “how does this help my mom?” type. Here’s a clue: Go watch one of your old Channel9 vids or one of your current PodTech vids, then immediately watch one of Rory’s Channel9 vids. You will see at once which interviewer has tech knowledge and which does not. Not that your interviews aren’t without merit, but they simply are not about technical details. So you can’t point to them as evidence that you actually know what you’re talking about at a technical level.

    Again, your whole point was that postponing PDC meant that Microsoft has nothing coming for developers this year when Orcas is on its way. Your entire premise is BS.

  98. “You’ve got to be kidding if you think the keynote with Mike Arrington and Ray Ozzie at Mix07 was any more developer focused than what Steve Jobs does at the WWDC.”

    Scoble, have you even bothered to watch the Mix07 keynote, which was mainly mc’ed by Scott Gutthrie, not Ozzie or Arrington. During the keynote, someone demoed actually creating and debugging a solution from start to finish. When has a WWDC *ever* showed debugging as part of a keynote? WWDCs are about, “look at this new Leopard feature!” nonsense that devs really don’t care about.

    “I’ve heard the .NET startup time problem from a range of developers across Microsoft’s Windows team. It isn’t a big deal to wait for Silverlight to startup, but it IS a huge deal to delay Windows starting up by a few seconds.”

    This is absurd. My own .NET apps start within 5 seconds if .NET hasn’t been loaded yet, and start instantaneously if it has been. And this has zero to do with waking from sleep mode. From where do you obtain your tech knowledge?

    And yes, I do know that you did LINQ interviews at Channel9. That doesn’t mean you actually know anything about the tech. Most of your questions are of the “how does this help my mom?” type. Here’s a clue: Go watch one of your old Channel9 vids or one of your current PodTech vids, then immediately watch one of Rory’s Channel9 vids. You will see at once which interviewer has tech knowledge and which does not. Not that your interviews aren’t without merit, but they simply are not about technical details. So you can’t point to them as evidence that you actually know what you’re talking about at a technical level.

    Again, your whole point was that postponing PDC meant that Microsoft has nothing coming for developers this year when Orcas is on its way. Your entire premise is BS.

  99. “Not even the games were written in .NET.”

    WTF?!?!?! What kind of excuse do they have for *this* one?

    ——-

    This is so tired. They don’t need an *excuse*. Why rewrite mindsweeper (for example) in .NET if they can leave the core as it is and just pretty-up the graphics?

    You don’t rewrite things just to rewrite them. For years, OSX’s Finder wasn’t written on top of Cocoa’s API. Did that mean that Cocoa sucked?

    Expression is written in .NET. That should be evidence enough that new apps are being written in .NET. Rewriting older apps like mindsweeper for no reason is stupid.

  100. “Not even the games were written in .NET.”

    WTF?!?!?! What kind of excuse do they have for *this* one?

    ——-

    This is so tired. They don’t need an *excuse*. Why rewrite mindsweeper (for example) in .NET if they can leave the core as it is and just pretty-up the graphics?

    You don’t rewrite things just to rewrite them. For years, OSX’s Finder wasn’t written on top of Cocoa’s API. Did that mean that Cocoa sucked?

    Expression is written in .NET. That should be evidence enough that new apps are being written in .NET. Rewriting older apps like mindsweeper for no reason is stupid.

  101. From my personal experience, .NET isn’t so slow that it would preclude loading it at boot time.

    I don’t have Vista, but I have XP with .NET 3.0 installed. Besides my own little side projects, the three .NET apps that I use regularly are Paint.NET, PowerShell, and XPS Viewer EP. Loading any of these before .NET has been loaded does take some time (8 seconds at the most). But once that’s done, loading any of them afterward takes little time, so you wouldn’t be able to tell the difference from a native app. But I have a relatively fast machine.

    Windows does take a long time to come out of sleep mode compared to OSX, but that has nothing to do with .NET. Windows XP is slow coming out of sleep mode even with no .NET libs installed at all.

    The reason Vista isn’t written in .NET is that Vista wasn’t rewritten from scratch. It has tons of old NT code in there (and rightly so; no need to change it other than recompiling it with the stack-checking flag).

    I’m not sure what any of this has to do with PDC’s postponment, BTW.

  102. From my personal experience, .NET isn’t so slow that it would preclude loading it at boot time.

    I don’t have Vista, but I have XP with .NET 3.0 installed. Besides my own little side projects, the three .NET apps that I use regularly are Paint.NET, PowerShell, and XPS Viewer EP. Loading any of these before .NET has been loaded does take some time (8 seconds at the most). But once that’s done, loading any of them afterward takes little time, so you wouldn’t be able to tell the difference from a native app. But I have a relatively fast machine.

    Windows does take a long time to come out of sleep mode compared to OSX, but that has nothing to do with .NET. Windows XP is slow coming out of sleep mode even with no .NET libs installed at all.

    The reason Vista isn’t written in .NET is that Vista wasn’t rewritten from scratch. It has tons of old NT code in there (and rightly so; no need to change it other than recompiling it with the stack-checking flag).

    I’m not sure what any of this has to do with PDC’s postponment, BTW.

  103. Dave: I was at Mix07 and watched the keynote. The Scott Guthrie part was excellent.
    I see you’re here just to throw insults my way. Congrats. Interesting that Channel 9 was built to 4.3 million unique visitors when I was doing videos at Channel 9 (I did most of the first 600 videos on Channel 9) and I don’t remember anyone complaining about our technical depth before I left Microsoft but now you are here saying it wasn’t deep enough. Sounds like sour grapes to me. You can’t complain about something two years after the fact and have any credibility.
    You’re a genius coder. Now go add something to Windows that we’ll all care about. Thanks.

  104. Dave: I was at Mix07 and watched the keynote. The Scott Guthrie part was excellent.
    I see you’re here just to throw insults my way. Congrats. Interesting that Channel 9 was built to 4.3 million unique visitors when I was doing videos at Channel 9 (I did most of the first 600 videos on Channel 9) and I don’t remember anyone complaining about our technical depth before I left Microsoft but now you are here saying it wasn’t deep enough. Sounds like sour grapes to me. You can’t complain about something two years after the fact and have any credibility.
    You’re a genius coder. Now go add something to Windows that we’ll all care about. Thanks.

  105. Kenny: there are large portions of Vista that WERE rewritten from scratch. The audio stack, for instance. Same with the networking stack. Yet the management forbade any team from working with .NET. Why? Because working on top of .NET would build dependencies on another team inside Microsoft. Also because .NET’s performance simply wasn’t up to the task.

  106. Kenny: there are large portions of Vista that WERE rewritten from scratch. The audio stack, for instance. Same with the networking stack. Yet the management forbade any team from working with .NET. Why? Because working on top of .NET would build dependencies on another team inside Microsoft. Also because .NET’s performance simply wasn’t up to the task.

  107. @57 “don’t remember anyone complaining about our technical depth before I left Microsoft but now you are here saying it wasn’t deep enough.”

    Uh…maybe because you were the only bar to be measured against? Lots of people thought the special effects in Star Wars in ’77 were cool, too. But compared to today? They sucked. Sort of the same thing. The tech content of your videos at Channel9 vs today’s Channel9? Well, I think Mr. Balder has a valid point.

  108. @57 “don’t remember anyone complaining about our technical depth before I left Microsoft but now you are here saying it wasn’t deep enough.”

    Uh…maybe because you were the only bar to be measured against? Lots of people thought the special effects in Star Wars in ’77 were cool, too. But compared to today? They sucked. Sort of the same thing. The tech content of your videos at Channel9 vs today’s Channel9? Well, I think Mr. Balder has a valid point.

  109. “there are large portions of Vista that WERE rewritten from scratch. The audio stack, for instance. Same with the networking stack.”

    Neither of which are candidates (nor examples of candidates) to be written on top of .NET. Corresponding .NET-provided services would be written on top of them instead. .NET is a high level framework, much of which sits on top of lowerlevel stuff (e.g. .NET’s internet transport api sits on top of wininet.dll).

    “Yet the management forbade any team from working with .NET. Why? Because working on top of .NET would build dependencies on another team inside Microsoft.
    Also because .NET’s performance simply wasn’t up to the task.”

    Well, at least the “dependencies” thing is true. Yes, the main problem with Vista’s delays was that lots of code was being written against other code (e.g. .NET 3.0) that was still in development, which resulted in code being checked in with no testing, leading to instable development process. The “reset” removed lots of these dependencies, many of which where .NET dependencies.

  110. “there are large portions of Vista that WERE rewritten from scratch. The audio stack, for instance. Same with the networking stack.”

    Neither of which are candidates (nor examples of candidates) to be written on top of .NET. Corresponding .NET-provided services would be written on top of them instead. .NET is a high level framework, much of which sits on top of lowerlevel stuff (e.g. .NET’s internet transport api sits on top of wininet.dll).

    “Yet the management forbade any team from working with .NET. Why? Because working on top of .NET would build dependencies on another team inside Microsoft.
    Also because .NET’s performance simply wasn’t up to the task.”

    Well, at least the “dependencies” thing is true. Yes, the main problem with Vista’s delays was that lots of code was being written against other code (e.g. .NET 3.0) that was still in development, which resulted in code being checked in with no testing, leading to instable development process. The “reset” removed lots of these dependencies, many of which where .NET dependencies.

  111. #59: fair enough. So, tell me, how much has Channel 9’s traffic gone up since I left? Compare that to the growth in traffic of ScobleShow.com. I have and ScobleShow has been growing a LOT faster and is quickly moving up the ranks of the Technorati list. ScobleShow reached top 2,000 blogs faster than Channel 9 did and Channel 9 had fewer blogs to compete against.

  112. #59: fair enough. So, tell me, how much has Channel 9’s traffic gone up since I left? Compare that to the growth in traffic of ScobleShow.com. I have and ScobleShow has been growing a LOT faster and is quickly moving up the ranks of the Technorati list. ScobleShow reached top 2,000 blogs faster than Channel 9 did and Channel 9 had fewer blogs to compete against.

  113. Scoble, RE: #44.

    My vista machine boots out of sleep mode before my monitor has an image on the screen, you can’t get much faster than that!

    I guess OSX’s speed is due to a few things, firstly the limited range of hardware it has to support would mean apple can tweak for that exact hardware and leave out everything else.
    Secondly, before the Intel switch the G4/G5 cpus meant that Apple *had* to optimize their software, whilst the PC camp was enjoying ever higher clock speeds and overall faster CPU’s so it was less of an issue – windows got faster by default.

  114. Scoble, RE: #44.

    My vista machine boots out of sleep mode before my monitor has an image on the screen, you can’t get much faster than that!

    I guess OSX’s speed is due to a few things, firstly the limited range of hardware it has to support would mean apple can tweak for that exact hardware and leave out everything else.
    Secondly, before the Intel switch the G4/G5 cpus meant that Apple *had* to optimize their software, whilst the PC camp was enjoying ever higher clock speeds and overall faster CPU’s so it was less of an issue – windows got faster by default.

  115. Will: well, I have a dual-core Sony Vaio which takes several seconds to get out of sleep mode while my new Mac gets out of sleep mode almost instantly. Maybe I should video tape the two coming out of sleep mode so you can see how much faster the Mac is.

  116. Will: well, I have a dual-core Sony Vaio which takes several seconds to get out of sleep mode while my new Mac gets out of sleep mode almost instantly. Maybe I should video tape the two coming out of sleep mode so you can see how much faster the Mac is.

  117. @63 – given you’re such a tech genius, I’m surprised you haven’t considered it could be the HARDWARE that is slow to wake up? My Dell takes less than a second.

    But I’m sure you’ll pull out some self-created stat to justify your all-knowingness. Just cos c9 had a bunch of hits in the early days is NOT a validation of the technical content level or quality. It’s probably all the moms you were asking questions for :-)

  118. @63 – given you’re such a tech genius, I’m surprised you haven’t considered it could be the HARDWARE that is slow to wake up? My Dell takes less than a second.

    But I’m sure you’ll pull out some self-created stat to justify your all-knowingness. Just cos c9 had a bunch of hits in the early days is NOT a validation of the technical content level or quality. It’s probably all the moms you were asking questions for :-)

  119. #64: congrats on getting a machine that wakes up so fast. I’ve seen dozens of Windows machines and most don’t wake up that fast.

    Regarding Channel 9. That’s not what people told us in the surveys that we did! But, interesting how the revisionists go back and make everything I did seem like shit. That’s cool. I guess if I was in the Microsoft ecosystem right now I’d be doing the same thing. Discredit Scoble’s work back then, cause that way we can discredit anything he says now.

    I understand how this game is played.

  120. #64: congrats on getting a machine that wakes up so fast. I’ve seen dozens of Windows machines and most don’t wake up that fast.

    Regarding Channel 9. That’s not what people told us in the surveys that we did! But, interesting how the revisionists go back and make everything I did seem like shit. That’s cool. I guess if I was in the Microsoft ecosystem right now I’d be doing the same thing. Discredit Scoble’s work back then, cause that way we can discredit anything he says now.

    I understand how this game is played.

  121. Scoble, I don’t doubt you, I’m just saying it is possible for windows to do the same, I guess it depends on the hardware. I’ll be swapping out my G4 Mac for a MacBook soonish, so I can see the difference then.

    Re: your channel 9 work. That was the reason I started going there, I really enjoyed your videos, and I enjoy your podtech videos too. Given the fact that was the main content on the site, I think your surveys were right ;)

    Would be nice to see some more discussions like the Vista/OSX video you did earlier in the year – I really enjoyed that.

  122. Scoble, I don’t doubt you, I’m just saying it is possible for windows to do the same, I guess it depends on the hardware. I’ll be swapping out my G4 Mac for a MacBook soonish, so I can see the difference then.

    Re: your channel 9 work. That was the reason I started going there, I really enjoyed your videos, and I enjoy your podtech videos too. Given the fact that was the main content on the site, I think your surveys were right ;)

    Would be nice to see some more discussions like the Vista/OSX video you did earlier in the year – I really enjoyed that.

  123. “Now that Google, Amazon, Apple, are shipping platforms that are more and more interesting to Microsoft’s developer community Microsoft has to play a different game.”

    This year, palmtop PC (UMPC) will become a real big business.
    I think Microsofts developer community should focus on this platform!
    We will really need palmtop-aware applications.
    :)
    http://www.itmedia.co.jp/news/articles/0705/16/news043.html

  124. “Now that Google, Amazon, Apple, are shipping platforms that are more and more interesting to Microsoft’s developer community Microsoft has to play a different game.”

    This year, palmtop PC (UMPC) will become a real big business.
    I think Microsofts developer community should focus on this platform!
    We will really need palmtop-aware applications.
    :)
    http://www.itmedia.co.jp/news/articles/0705/16/news043.html

  125. There are plenty of reasons to encourage .NET for reasons within the Windows ecosystem, that have nothing to do with competitors to Windows that Microsoft is ‘trying to destroy’. One thing is .NET code is safer than what the average programmer would do, one is that if all major applications are on a virtual layer with .NET it will be easier to change the core of Windows without breaking things, one is that they can encourage different practices, including like parallel computing, by building .NET libraries around it, one is that it makes the system potentially more object-oriented and accessible to object-oriented systems, one is that there are advantages to WPF and other WinFX pillars that will make more attractive applications for Windows,… well, so, the point is people who hate Microsoft assume that there’s no reason at all to have a coherent, integrated desktop platform, and any attempt to promote it is really on an ulterior level an attempt to destroy competitors because of proprietary tie-in. Sorry, Microsoft-haters, you really don’t have the whole story.

  126. There are plenty of reasons to encourage .NET for reasons within the Windows ecosystem, that have nothing to do with competitors to Windows that Microsoft is ‘trying to destroy’. One thing is .NET code is safer than what the average programmer would do, one is that if all major applications are on a virtual layer with .NET it will be easier to change the core of Windows without breaking things, one is that they can encourage different practices, including like parallel computing, by building .NET libraries around it, one is that it makes the system potentially more object-oriented and accessible to object-oriented systems, one is that there are advantages to WPF and other WinFX pillars that will make more attractive applications for Windows,… well, so, the point is people who hate Microsoft assume that there’s no reason at all to have a coherent, integrated desktop platform, and any attempt to promote it is really on an ulterior level an attempt to destroy competitors because of proprietary tie-in. Sorry, Microsoft-haters, you really don’t have the whole story.

  127. “This is a bummer, because .NET is a lot easier to write than C++ and letting Microsoft’s developers write .NET code for Windows would unleash a bunch of innovation.”

    I’m pretty sure the core of Windows is C there Robert, not C++. You see, even C++ is not performant enough to write schedulers and low level IO.
    .NET makes writing very high level apps easy, it does not make writing system code easy at all. It makes it harder.
    Where you could just manipulate memory, now you have to go through several levels of indirection with .NET or Java.

    I can understand how Robert is confused only having a very high level understanding of programming and technology though.

    I would suggest he buy a little AVR starter kit:
    http://www.smileymicros.com/
    and
    http://www.avrfreaks.net/

    for further info. Once he does the little “kid level” kit, his posts will snap into correctness. It’s worth it Scoble.

  128. “This is a bummer, because .NET is a lot easier to write than C++ and letting Microsoft’s developers write .NET code for Windows would unleash a bunch of innovation.”

    I’m pretty sure the core of Windows is C there Robert, not C++. You see, even C++ is not performant enough to write schedulers and low level IO.
    .NET makes writing very high level apps easy, it does not make writing system code easy at all. It makes it harder.
    Where you could just manipulate memory, now you have to go through several levels of indirection with .NET or Java.

    I can understand how Robert is confused only having a very high level understanding of programming and technology though.

    I would suggest he buy a little AVR starter kit:
    http://www.smileymicros.com/
    and
    http://www.avrfreaks.net/

    for further info. Once he does the little “kid level” kit, his posts will snap into correctness. It’s worth it Scoble.

  129. Robert,

    1. Powershell is .NET code and it is a very important piece of Windows Server 2008. Also, I think IIS 7 is so entrenched with ASP.NET you cannot tell where one starts and the other ends (but it is not necessarily a lot of .NET code, this I don’t know). All this is server stuff, but guess what? The same happened with Java. The sweet spot for managed code is on the server. On the client it can become as good as on the server once you overcome some limitations, like the loading time problem you mentioned. Anyway, .NET is very usable on the client side and I expect Microsoft will use it more and more in its own products. I think there is some kind of culture clash here.
    2. For those that argue about the limitation to one CLR for process, I heard that future versions of .NET will run on only one CLR (the latest installed), only the BCL will differ, depending on the one compiled your apps against. This is the aim of the red / blue separation of code in the framework. Also, Scott Guthrie says that you can load Silverlight and other version of the CLR on the same process.
    3. Speaking of Silverlight, I don’t get what you mean by “it is a repurpose of .NET and it is not major”. It is indeed major (and it took a lot of work for Microsoft), and the fact that it is still .NET is its killer feature. Don’t forget the million of .NET developers that are happy to be “repurposed” along with it.
    4. About the PDCs being about Office and Windows as development platforms, I don’t think those are much relevant. The primary Microsoft “development platforms” these days are .NET and SQL Server, Windows and Office Server like Sharepoint are third and fourth, and Windows Live is probably on the rise. The Longhorn PDCs were very exciting because they were about the future of .NET and Microsoft eating its own dog food, making .NET even better. When they decided that they would not build Vista on WinFX but would instead backport most of WinFX to XP nobody cried (well, maybe some cried for WinFS). .NET got better anyway.

    Regards!

  130. Robert,

    1. Powershell is .NET code and it is a very important piece of Windows Server 2008. Also, I think IIS 7 is so entrenched with ASP.NET you cannot tell where one starts and the other ends (but it is not necessarily a lot of .NET code, this I don’t know). All this is server stuff, but guess what? The same happened with Java. The sweet spot for managed code is on the server. On the client it can become as good as on the server once you overcome some limitations, like the loading time problem you mentioned. Anyway, .NET is very usable on the client side and I expect Microsoft will use it more and more in its own products. I think there is some kind of culture clash here.
    2. For those that argue about the limitation to one CLR for process, I heard that future versions of .NET will run on only one CLR (the latest installed), only the BCL will differ, depending on the one compiled your apps against. This is the aim of the red / blue separation of code in the framework. Also, Scott Guthrie says that you can load Silverlight and other version of the CLR on the same process.
    3. Speaking of Silverlight, I don’t get what you mean by “it is a repurpose of .NET and it is not major”. It is indeed major (and it took a lot of work for Microsoft), and the fact that it is still .NET is its killer feature. Don’t forget the million of .NET developers that are happy to be “repurposed” along with it.
    4. About the PDCs being about Office and Windows as development platforms, I don’t think those are much relevant. The primary Microsoft “development platforms” these days are .NET and SQL Server, Windows and Office Server like Sharepoint are third and fourth, and Windows Live is probably on the rise. The Longhorn PDCs were very exciting because they were about the future of .NET and Microsoft eating its own dog food, making .NET even better. When they decided that they would not build Vista on WinFX but would instead backport most of WinFX to XP nobody cried (well, maybe some cried for WinFS). .NET got better anyway.

    Regards!

  131. @61. Wait! You’re going back and checking C9 stats today compared to what they were when you were there? Then you are tracking that compared to your current stats? Can you say “insecure”? I thought you could.

    Just because it had a lot of hits isn’t a indication of quality. Hell, more people read People or USA Today than the NY Times or The Economist. Does that mean People and USA Today have better, more informative content?

  132. @61. Wait! You’re going back and checking C9 stats today compared to what they were when you were there? Then you are tracking that compared to your current stats? Can you say “insecure”? I thought you could.

    Just because it had a lot of hits isn’t a indication of quality. Hell, more people read People or USA Today than the NY Times or The Economist. Does that mean People and USA Today have better, more informative content?

  133. “This is a bummer, because .NET is a lot easier to write than C++ and letting Microsoft’s developers write .NET code for Windows would unleash a bunch of innovation.”

    Robert, that statement is clinically retarded. That’s like saying writing a novel in French instead of English will unleash some kind of new theme in literature.

  134. “This is a bummer, because .NET is a lot easier to write than C++ and letting Microsoft’s developers write .NET code for Windows would unleash a bunch of innovation.”

    Robert, that statement is clinically retarded. That’s like saying writing a novel in French instead of English will unleash some kind of new theme in literature.

  135. David Balder: I’m not talking about Minesweeper. In case you haven’t noticed, Vista has a lot more built-in games than XP.

  136. David Balder: I’m not talking about Minesweeper. In case you haven’t noticed, Vista has a lot more built-in games than XP.

  137. “Apple’s WWDCs, which, despite the “DC” in the name, are aimed at consumers rather than devs. ”

    What are you, stupid or something? You’ve obviously never attended Apple’s developer conference. Take a look at the session descriptions, and then try to tell me that WWDC is for consumers.

  138. “Apple’s WWDCs, which, despite the “DC” in the name, are aimed at consumers rather than devs. ”

    What are you, stupid or something? You’ve obviously never attended Apple’s developer conference. Take a look at the session descriptions, and then try to tell me that WWDC is for consumers.

  139. “For years, OSX’s Finder wasn’t written on top of Cocoa’s API.”

    It still isn’t.

    ” Did that mean that Cocoa sucked?”

    No, it means that the Finder sucks.

    Seriously, the Finder is the saddest example of political compromises in Apple’s entirety of software offerings. I hope they’ll finally take it out behind the barn and shoot it in time for leopard, but I’m not holding my breath.

  140. “For years, OSX’s Finder wasn’t written on top of Cocoa’s API.”

    It still isn’t.

    ” Did that mean that Cocoa sucked?”

    No, it means that the Finder sucks.

    Seriously, the Finder is the saddest example of political compromises in Apple’s entirety of software offerings. I hope they’ll finally take it out behind the barn and shoot it in time for leopard, but I’m not holding my breath.

  141. @Diego – You are talking about things (red/blue) which weren’t even planned properly when Longhorn was heading into beta. If you know anything about what happened for the Longhorn reset (I do and most commentators I’ve seen do not) you would know there were both political and technical reasons for the mandate against managed code.

    Incidentally, Robert is completely correct that startup times (working set goals not being reached, among other things) and versioning (servicing and of the CLR and of libraries, especially for platform features) were key technical issues, though there were others. It is also true that managed code itself was not always the main problem, and that many solutions to the problems had been considered.

    @.Net versus C/C++ – this is a religious debate to some extent but it is very true that C# managed code development of libraries is of higher quality and quicker turnaround than C++ development. I would say undeniable but that is too strong. This is not just for buffer overruns and memory leaks, though those are the poster children. Debuggability and code analysis, in fact maintenance in general is much easier and quicker for managed code.

    There is a cost for using managed code however.

    @Robert – I do agree with your overall sentiment that Microsoft should be learning from the embarassments of the Hailstorm and WinFS announcements / followed by cuts, for example. I do think that comparing Microsoft to Facebook and finding it wanting is kind of an overreach on your part, though you are I guess narrowly focused now on the Web 2.0 development sphere or whatever you hip folk are calling it now :)

  142. @Diego – You are talking about things (red/blue) which weren’t even planned properly when Longhorn was heading into beta. If you know anything about what happened for the Longhorn reset (I do and most commentators I’ve seen do not) you would know there were both political and technical reasons for the mandate against managed code.

    Incidentally, Robert is completely correct that startup times (working set goals not being reached, among other things) and versioning (servicing and of the CLR and of libraries, especially for platform features) were key technical issues, though there were others. It is also true that managed code itself was not always the main problem, and that many solutions to the problems had been considered.

    @.Net versus C/C++ – this is a religious debate to some extent but it is very true that C# managed code development of libraries is of higher quality and quicker turnaround than C++ development. I would say undeniable but that is too strong. This is not just for buffer overruns and memory leaks, though those are the poster children. Debuggability and code analysis, in fact maintenance in general is much easier and quicker for managed code.

    There is a cost for using managed code however.

    @Robert – I do agree with your overall sentiment that Microsoft should be learning from the embarassments of the Hailstorm and WinFS announcements / followed by cuts, for example. I do think that comparing Microsoft to Facebook and finding it wanting is kind of an overreach on your part, though you are I guess narrowly focused now on the Web 2.0 development sphere or whatever you hip folk are calling it now :)

  143. […] Duncan Riley wrote an interesting post today onHere’s a quick excerptThe person who told me this (who works at Microsoft) told me .NET still takes too long to startup and load into memory and because Windows is now being compared to OSX they can’t afford to ship components that would slow down Windows. … […]