WinFX: Now Better than Ever

By now you have heard the news: WinFX will be shipping on XP, WS03 in order to make both Longhorn and WinFX ship with higher quality and sooner.  If you haven’t already, check out JimAll’s Channel9 interview.  Longhorn will contain core-OS level features focusing on the basics: security, reliability, manageability and performance (and cool new look-and-feel feaures). While WinFX will be shipped with Longhorn and available for XP SP2 and WS03 in order to expose the developer platform to a broader reach.  Of course Longhorn will ship the .NET Framework 2.0 (Whidbey) redist and WinFX, but there will not be lots of interdependencies dependencies between WinFX and Longhorn.  Along those lines, WinFS, the storage engine for Windows, will not be on this train.  It will likely be in beta when LH RTMs.


At PDC2003 Jim Allchin announced WinFX.  At the time he made a passing comment that really stuck with me.  I imagine that most folks missed it or quickly forgot it.  But he clearly said that this is an unprecedented early-look at a major OS release and that some of what you are hearing today will NOT make it into the final product.  Turns out he knew what he was talking about (a bazillion years shipping Windows releases will do that for you I guess ;-)).   While it is true that some of what we talked about at that PDC will not make it into the LH product, I still strongly believe in the vision we laid out and the company is still executing on that vision, it is just going to take us a couple of hops to get all the way there. 


I think this is a great move for a couple of reasons:

1. Reach.  We are now able to bring the greatness of WinFX to a much broader set of customers, making WinFX an even more compelling development platform.

2. Timing. This plan optimizes for shipping sooner.  And that is goodness for the entire ecosystem.  There is really only one way to make a software project ship sooner: Simplify.  And simplify is what we did with Longhorn and WinFX.  We simplified the features, simplified the dependencies, we simplified the product.  As an aside, adding resources rarely makes the project more simple and therefore often does not help you ship sooner. 


As far as the stuff I deal with on a day-to-day biases things will get a lot easier for me.  We have a more scoped set of APIs in WinFX so I can focus better on cleaning those APIs up and making them shine.


What do you think? Good move or not?  


If you had $100 to spend today on everything you ever heard would be in “WinFX” or “Longhorn” where would you spend it?  This is exactly the question we are asking internally now as we crisp up our definition of WinFX over the next few months.  Your feedback would be very valuable. 

Comments (25)
  1. Dono says:

    "What do you think? Good move or not?"

    I see both good and bad in it.

    Good: For the many people who have XP now, this is great news. I think it is great that more and more .Net technologies will be availble to many.

    Bad: However, if Avalon, Indigo, and WinFX are going to be available for XP, then what reason is there to consider upgrading to Longhorn? I think that Microsoft is hurting their future OS potential. While it is early, I think that Microsoft needs to clarify what Longhorn will offer over the enhanced XP of 2006.

  2. Tom Kirby-Green says:

    Yes Avalon will be on XP and 2K3, but Aero the shell won’t. If the new shell is groovy enough then I think enough people will upgrade – what concerns me slightly is to what degree was Aero’s coolness infact down to WinFS. However I do think you have a point, put Avalon on XP and folks like Stardock and other writers of alternative XP shells today will be competing to out-do Microsoft at their own game.

    The uber-brains at MS must have thought about this since they need to ship a lot of copies of Windows 2007 (or whatever Longhorn ends up being called) to keep shareholders happy.

    That said I can’t help feeling that the real winners here are nVidia and ATI.

    When Brad are we going to see a developer release of WinFX for downstream platforms? And will there be immediate support for X64 (can’t wait to run this still on my Nocona box)?

  3. sbjorg says:

    I would assign the $50 to track down the individuals who over-promised features and hid the real risks from management to give them to boot. And the remaining $50 would go towards cloning Raymond Chen.

  4. Slavo Furman says:

    Anyone knowns if MSH (Microsoft Shell, Monad) will be also available for WinXP and Win2003?


  5. Kevin Daly says:

    This is pretty much what we’ve understood the situation to be with Indigo for some time.

    As far as Avalon goes, I don’t think having it available for XP will hurt Longhorn sales at all: to get the full experience you will most likely need new hardware, and new machines will ship with Longhorn…plus I wouldn’t be surprised if even aside from the hardware the optimal experience is affected by cored OS-level features. In addition, the broad availability of Avalon provides a much greater incentive for developers to er, develop for it – and for designers to start targetting XAML (presumably using nice friendly IDE’s to do so). This will actually accelerate a process I’ve been expecting ever since the PDC, namely the evolution of today’s web designers into more generalised UI designers.

    The WinFS news is the disappointing one of course (well, that and Indigo not being available *before* Longhorn). I hope it’s not too far in the future, and that I get to use it while I still have some hair.

  6. Ovidiu says:

    It seems to me that Microsoft is losing on the PR front. Announcing features and then cutting them back is just bad for your own image – imagine one of your teammates saying "I can do it… I can do it…" and three months later: "I gotta cut this feature, or I’ll fall way behind schedule". Do you kill him or do you kill him?

    Also, I think Longhorn is losing selling points. Plenty of people are already sticking with whatever is good enough for them, be it Windows 98/Me or NT4/Win2k. Longhorn will sell even less than Windows XP (that hit record lows in its time afaik).

    Backporting features to previous versions of Windows might mean that advanced characteristics won’t be available on those platforms OR that MS will have do do extra work to make it happen (like the extra work done in the CLR to make it happen on Windows 98) and I don’t know how wise an investment that is. You’ll end up writing service packs for Windows 2000, XP and 2003 forever.

    On the other hand, since the CIO guy is gone and the motto is "Do more with less", this might just make happy the premium customers that have large numbers of existing Windows versions installed and that do not have to upgrade (yet).

    One last set of observations: "Longhorn will contain core-OS level features focusing on the basics: security, reliability, manageability and performance (and cool new look-and-feel feaures)." is not believable anymore for two reasons:

    First, security appears in the head of the list, but the overall quality of some Windows components is low (or perceived as such). Of course, I mean the browser, the scripting engine and the (unsecure) integration between these and a few applications that forces inappropriate responses (such as application level security that would belong in the scripting engine for Office).

    Second, I really don’t care about the cool new look-and-feel features. Most of them are marketing hype. I don’t know how many people buy into the "hip" factor.

    I guess you already figured, if you give me WinFX I’ll just use it and forget about Longhorn 🙂

    Hope this helps.

  7. I’d spend the $100 making sure you get the shared GPU driver model into WinFX on XP and Windows 2003. I know you need something to sell Longhorn but if WinFX on XP is a subset of the full WinFX no one is going to write apps that take advantage of the full Longhorn feature set.

  8. Tom asks when we are going to ship WinFX…. Our goal is to ship it with Longhorn, so we are targeting locking down a few months before LH RTM and shipping on the same day.

    Tom also asks about X64 – While I still haven’t got my x64 machine on my desktop yet ;-), we are still targeting x64 to ship at the same time as x86. BTW, Whidbey will support X64 even sooner!

  9. David says:

    Ok, I am going around preaching this on every blog, I think: PLEASE call it .Net 3.0! Drop the WinFX name. ONCE show sanity with your naming. Remember "ActiveX"? Right…

    WinFX made a lot of sense to differentiate the stuff that was only available on Longhorn and not on downlevel operating systems. But now, it would really make things so much more logical and cause a lot less confusion if Avalon and Indigo would just be packed together with the Framework as .Net 3.

    Please 🙂

  10. Tom Kirby-Green says:

    Yeah, "WinFX" is pretty terrible, let’s hope its a placeholder like "Longhorn". "ActiveX" was a disaster, I still field questions from people asking how ActiveX differers from COM differs from OLE. On the other hand Microsoft does need something to bridge Win 32 / Win 64, especially as we transition from one to the other. It’s hard to come up with names that don’t upset someone, what about WinNG? (Next Generation), or WinPK (Penguin Killer). Only serious.

    Getting back to WinFX and downstream platforms, can we expect the first WinFX ‘XP/2K3 SDK beta along with the Longhorn beta 1? Will Visual Studio Orca fully support WinFX on ‘XP etc, in particular will all the designers and life cycle tools support downstream platforms for WinFX?

  11. theCoach says:

    The above is brutal to your credibility (and mine, as I have used that post to communicate that cutting WinFS was being over-hyped). Of course, things happen and you re-evaluate, but the above post looks VERY bad.

    WinFS was to me the heart and soul of Longhorn, and I am looking forward to it when the Server version is released. It is more important that you get it right, but I might suggest, if you are able, to come up with some guidance for application data being stored in a way that will migrate to WinFS. Perhaps Indigo will encourage more collaboration between local applications, but it was my understanding that WinFS had the potential for building domain-specific data that could be used across applications, with schemas evolving for different spaces.

  12. Chui Tey says:

    In the end, for each person there is one compelling feature that makes upgrades worthwhile.

    In Windows 3.0, it was making the dot matrix printers print better than they’d otherwise do.

    In Windows 95 it was plug and play.

    Windows 2000 was its unprecendented stability.

    Windows XP was an improved shell.

    As great as WinFX is, it’s a brand new platform. Even windows 2000 can’t run WinFX programs, and that’s most of the corporate PCs. Then you need a high-end graphics card, well that’s not standard issue either. Compelling as it is, it depends on us buying new hardware before it becomes possible. Application developers are going to find it easier to develop for the Mac before targeting the WinFX market.

  13. Since it comes up a bit:

    What’s the document that most accurately describes *exactly* what WinFX is? What API’s are included?

    Is it by definition whatever .NET framework ships as part of Longhorn?

    It seems harder to get people excited about something that has such a shifty meaning – Avalon, Indigo, and WinFS are all very clear in at least what their intentions are, even for developers not clear on the actual API. WinFX seems to have no semantic associated with it "in the wild" at the moment AFAICT.

  14. Judah Himango says:

    "What do you think? Good move or not? "

    We lose WinFS in return for a nearer ship date and an API that will work on XP? Good move, hands down.

  15. Laura T. says:

    I’m quite puzzled. I’m not sure if this is a winning move for Microsoft.

    I’m glad that the WinFX and others will be ported to XP and 2003, but I’m quite unhappy that this means more and more incompatible PCs. Not only you must stuff >20MB of WinFX but also Indigo and Avalon into your distribution CD. You must now check what is present, what’s not, what level they are (there will be SPs and other things).. the horror of IE versions are coming back. With Longhorn you could be sure about these things.

    And of course XP and 2003 will get bigger, much bigger and slower than they are already. I am still very unhappy with XP, 2003 is better though..

    A good point is that the migration curve will be much lower. Longhorn will not be a completely new world. That’s why it is a good thing to remove WinFS too.. it would have been really too much to handle for ISVs.

    Let’s face it, it’s not so much OS world anymore. The application layer, WinFX, Indigo and Avalon are much more important to the rest of the world, than core OS tweaks (OK, Longhorn would not be a little tweak, but still..). We need better applications, and a good OS (like w2k), but a lot better application support, WinFX, Avalon and Indigo. Just make sure that the versions of WinFX, Avalon, Indigo et. al. are the same in Longhorn (no more, IsSupported(xxx) functions ).

    What I hope, that there will not be any intermediate OSs between XP/2003 and Longhorn. The Win9X/SE/SE2/ME nightmare must not resurrect.

    So, actually I’m happy. I’ll get better application services and infrastructure to make better apps sooner. Without a BIG learning curve. Just don’t cut much more.


  16. mikeb says:

    To Slavo Furman: the information I got from WinHEC 2004 was that Monad will be supported on Windows 2000, Windows XP, Windows Server 2003, and subsequent versions of Windows.

    To BradA: FWIW, I agree with David… rename WinFX to .NET 3.0.

  17. Kevin Daly says:

    On the WinFX rename issue: I’m not sure that .NET 3.0 is such a good choice, since WinFX is a much larger slice of the API pie than previous .NET incarnations, so calling it .NET 3.0 might suggest a more incremental change than is actually involved.

    On the other hand, WinFX as a name stinks big time, if only because it sounds too much like WinFS (making phone conversations on the subject next to impossible).

    "FX" doesn’t scream ".NET" to most people as it apparently does to Microsoft staff.

    On the other hand I can’t think of any decent acronyms (for example, "MAW" anybody? ("Managed API for Windows") …And Lo, I Have Plumbed The Depths Of Lameness).

  18. Shaun Newman says:

    I am pleased that Windows XP and WS03 will still have shelf life for a while, it gives us a chance ti weigh up the options and prepare for Longhorn. However I am sceptical, Windows XP gave us a fair bit over previous OS’ and was a welcome relief bringing stability and security, and a great deal of hope after ME.

    While features are being tweaked and trimmed from Longhorn I hope there will be no major compromise in the overall vision and aim. I remember reading that Longhorn was to re-write the rule book when it comes to the OS.

    Shareholders and greedy users like be asside, we are notorious in this business for settling for less in order to meet some deadline. While understandable in a cut-throat world where time is money, the real revolution would be to spend the time and get it right.

    Now, if we can just find the cure for scope-creep…

  19. Last year, Chris Sells wrote a series of articles in which he set out to rewrite Solitaire using Avalon. …

  20. Last year, Chris Sells wrote a series of articles in which he set out to rewrite Solitaire using Avalon. …

  21. Brad Abrams : WinFX: Now Better than Ever.

    The comment is from “David” who states”…

Comments are closed.

Skip to main content