Ballmer’s .NET visions. They’ve taken 13 years, but we’re living them today

Ballmer’s .NET visions. They’ve taken 13 years, but we’re living them today

Summary: Microsoft's 2000 concept videos: were they a success or a failure? Today's IT landscape suggests they succeeded - and beyond Microsoft's wildest dreams.


Way back when, in the spring of 2000, was Microsoft’s Forum 2000 event where it unveiled .NET and where it showed its vision for the future of computing in a series of videos.

Those videos are the thing most people remember from the event, classic examples of what Bruce Sterling calls "design fictions", stories designed to inspire and direct development.

13-09-2013 13-38-55
From Microsoft's Forum 2000 concept videos, a healthcare dashboard - much like the one in my doctor's surgery today.

So perhaps it wasn’t surprising to read a recent article that stated Microsoft hadn't delivered on the promise of the videos. Except the videos weren’t a promise; they were a what-if, a "look what we could build", videos that set the scene for a vision of a very different world of computing from the client-servers that powered business and the early internet.

Thirteen years into that future and it's clear that, while not every detail has arrived just the way the videos depicted, we are living in that world. We're using service-oriented architectures to connect disparate business and technologies, without having to worry about whether one is Java, if one is .NET, if one is a PHP web app. We’re storing our lives in cloud services, using them to mediate our records, our connections, the very devices we use.

People often think of Microsoft as wedded to the vision of the PC on the desk, but it's clear from those vision videos that the Microsoft of 2000 was already a very "post-PC" organisation, with a view of the future that was focused on cloud services and mobility. Sure, there have been missteps on the way.

We should also remember what Microsoft Research’s Bill Buxton calls the "Long Nose" of technology — the 10 years and more that it takes to leave the research lab and get into the hands of the consumer. It's an effect I have personal experience of, working in a research lab in the early 90s and experimenting with early versions of cable modems, of ADSL networks, and of pocket sized screens and small radios. At the bleeding edge then, they're in most homes today — costing a fraction of the thousands of dollars my experimental networks cost to build.

And so, we should look at the technologies we have today, where iPhones with biosensors unlock information in the cloud, where a Windows 8.1 PC shares apps with the rest of the machines you own, and where the same code runs on the desktop, on the phone, and in the cloud. Watch the videos again, and you’ll see much of the technology landscape we live in. it may not all be here yet, and it certainly won't look it does in the video, but that Forum 2000 vision (like Apple’s Knowledge Navigator before it) is part of the foundations of today’s IT world.

13-09-2013 13-40-37
Apple's Knowledge Navigator: design fiction from 1987 that arrived twenty years later as Siri and the iPad

Sure, there's still work to be done. But the world we live in is a different one from that spring day. There's a joke about a man in a car, lost somewhere rural. He sees a farmer in a field, and stops his car to ask how to get to his destination. "Do you know the way?" he says. "Of course," replies the farmer, "But I wouldn't start from here."

Looking at those videos now, and it's clear we've ended up somewhere different, and if we were to build them today we wouldn't be starting where they started. But we have most of what they talked about, and an infrastructure that would build the rest — if they were what we actually wanted now. We have a world where pocket computing devices link us into global communities, where we can pick up new hardware and in minutes get access to our applications and data, where we can control who can see what about us, where electronic records and open APIs have enabled new workflows, new ways of information sharing.

Back in 2000, we'd have looked at what we have today, where we treat smartphones as cognitive prostheses, and where we live lives as much on as offline, as a design fiction.

If someone had made a scenario video of our daily lives, we'd have probably giggled a bit, and then gone off to try and build the bits we thought were cool — and ended up building something that had some of those ideas at its core, but was actually more useful. That’s what Microsoft’s engineers did with those videos — and not just Microsoft’s, engineers who left and went elsewhere, and engineers at other companies.

And because they did, the videos didn't fail. They succeeded, making all our lives better.

Further reading

Topics: Steve Ballmer: The Exit Interview, Microsoft

Simon Bisson

About Simon Bisson

Simon Bisson is a freelance technology journalist. He specialises in architecture and enterprise IT. He ran one of the UK's first national ISPs and moved to writing around the time of the collapse of the first dotcom boom. He still writes code.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Yep

    .NET, with its robust support for SOAP and Internet connected communications (HttpWebRequest and XmlReader/XmlDocument), wasn't built purely for the desktop. It was built from the ground up as a cloud connected development environment. It was a refreshingly firewall friendly technology, as opposed to the weird complexity of JavaBeans and CORBA.

    Things have changed, and the complexity of SOAP is being displaced by RESTful services... but the basic ideas are the same. They just got their sooner.
    • Not only that

      If you look into the CLR internals they are highly optimized for single instance deployment unlike needing runtime for each app running in competing technology. The .NET is also optimized for platform portability.
      Ram U
    • @ Mac_PC_FenceSitter

      I would assess that SOA has actually failed to get adopted by enterprise or internet or government. SOA principles are partially reused within RESTful architecture but the underlying problems of REST resource maintenance including statelessness about client prevent it being truly a SOA Web Service. Some call REST a Web API (as .NET itself calls it) and that is all it is. It is not really a SOA architecture unlike SOAP or other protocols including Java RMI or DCOM.

      REST sells since it is based on HTTP protocol for transport. Other than that, using URIs for resources and building methods around them simply does not constitute SOA.
      • I'm not aware of any deficiency in SOA getting adopted

        Microsoft's approach was only tangentially stateful (WCF could hold state, but SOAP was never designed to.) SOAP is certainly widely used... and REST certainly, in my view, is an SOA approach... it just relies on the client to hold state, like SOAP before it.

        I don't view RMI or DCOM as SOA... distributed component architecture, intermachine communication, but not true SOA, which presumes that services are hosted on computers at least somewhat optimized for the purpose.
      • Big difference

        SOA is architecture SOA service could be running using SOAP, ReST over HTTP or any other service. if your intention is SOAP then it is different. A lot of organizations are into SOA whether intentionally or not.
        Ram U
  • Those videos

    Are about what they could have done, but failed in many ways.
    Conceptually they can be interesting, as many things were in "Space 1999" (note the TV series is from mid 70s not end of 20 century).
    • Videos should have been burned

      The fact is many people have had more or less accurate visions and most have failed between vision and execution. Most organizations fail in executing their vision, there is nothing new about that, it's quite common. The rare organization that is successful in executing its vision is typically rockin it in the industry unless your vision is to plod along in mediocrity never rocking the boat or drawing attention.

      Another bad decision, rather than open up for certain criticism they should have just burned those videos.
  • Too bad they screwed up Windows 8 by going back to COM...

    But that's what happens when you put worm ridden filth like Steven Sinofky in a management position.
    • Steve not invited to your house for Thanksgiving then?

    • How did Microsoft "go back to COM"?

      What they, in fact, did was bring COM forward into the .NET world.

      Yes, the new WinRT API's expose a set of native COM objects via CLR-compatible metadata and language projections making integration between .NET and the new WinRT API's seamless and blazingly fast. It also allows JavaScript code to access these new API's quickly and naturally.

      One of the benefits of building these new WinRT API's as COM objects is that they don't require .NET in order to run. This means that if you have a pure C++ app, it can access these same API's without the memory and IO overhead of loading, initializing and hosting the CLR & .NET within the app's process.

      Had Microsoft implemented the WinRT API's in .NET, every process that wanted to call those API's would have had to load, initialize and host the CLR and invoke those API's through an interop layer. For many types of app this would be less than optimal, if not impossible!
      • please

        A duck in a chimp suit is still just a duck, language projections or no. All that idispatch Marshall by ugliness is still right there, under the skirts, in all its unmanaged guid ID glory.

        Sinofsky and his WinDiv followers won I guess. Developers lost.
      • Right, part of the Longhorn debacle was the effort to put .NET into an OS

        It just doesn't fit there - it really wasn't designed to be used that close to the metal. That's why WinRT uses services that are OS-capable-proven. However by making it transparently available to .NET clients (as well as C++ and JavaScript clients) you end up with something that satisfies more needs.

        It's certainly a lot more efficient that the original P/Invoke interop model.
  • I Don't Know

    It seems that more emphasis is placed on the altruistic motives of Microsoft, when I would call those pronouncements the sales pitch.

    Clearly everything has expanded in the client/server world. .net was the alternative to java, but, from my perspective, more and more people stopped asking the question that was going to be answered by java EE. Search, MapReduce, BigTable, NoSQL, Hadoop, Ruby on Rails, the rise and fall (in buzz) of Drupal and the other CMS, etc., these are all things that emerged in the last 13 years and, in some respects, they were routing around Sun and Microsoft. (Though Hadoop is built on java.)

    The language nerds — I say that with love and as a language hobbyist — are really excited over Clojure and Scala, both of these built on the jvm, and both have auxiliary branches that utilize the clr, but the language leaders are not prioritizing ease of porting. Does that conform to Microsoft's vision of 2000? IronPython and IronRuby being activated and then jettisoned suggests that Plan A and Plan B didn't quite work out as expected. Microsoft is still providing a jdbc driver for SQLServer; that certainly has to be a resignation to the world making other choices, despite the power of .net.

    Sun imploded, but IBM's embrace of Linux and pivot to emphasizing services may have been more key than the truly successful achievements of Microsoft's competition built on Intel processors.

    Still, it looked to me back then that Microsoft desire to kill java — first expressed as J++ — so as to attack a vulnerability for both Sun and IBM, went unfulfilled. Did Sun's endorsement of java on Linux and BSD and license under GPLV2 and endorsement of OpenJDK mean that java would always be a first choice for people who wanted to explore ideas without plunking down for a Windows server license? I would guess so, and I think the industry undervalues what the kids starting from scratch on low-cost/free as in beer platforms can do, until they become Google or Facebook. This guy found mono to be unreliable in that sometimes it would build and sometimes not. I stopped caring and trying to make it work. Microsoft's endorsement of Ximian was rather half-hearted: blame the lawyers for putting precious ip in a vault.

    Here's the point I'd make, Microsoft's accomplishments with .net are real, but their 2000 ambitions for the platform were far from realized. It was a defensive move. As for Ballmer's foresight in 2000, nope, not convinced. And let's be frank, .net got really interesting when Microsoft Research hired the Haskell guys who made the case that FP was an important thing to get into .net. V2 had infrastructure builds and V3 started incorporating lambda and first-class functions into C#. In other words, C# started being something else other than a quibble with Gosling regarding boxing and other arcana.

    I am fairly sure that as the roadmap was being laid out for Ballmer, he raised his hands and said "Stop, you had me at java-killer." I speculate, but I stand by my assertion that that is the yin and yang of Ballmer: open-throttle, engines-ahead-full competitor to the obvious.
    • I don't know why so many people hype themselves up over Big Data

      The number of organizations outside the Fortune 500 who need big data could be counted on two hands. Four maybe.

      As far as I am concerned, outside Social Networking plays, businesses that play buzzword bingo and get themselves tied up in fashionable approaches are going to end up spending four times as much on their development efforts. Java EE, PHP and .NET tied to a MySQL/SQLServer db do solve some neat problems, and for most every day purposes do a much neater job of it than NoSQL, Hadoop, etc.
  • None Of Which Has Anything To Do With Dotnet

    Even Microsoft has moved away from Dotnet and its retarded nephew, Silverlight, in favour of WinRT as its new architecture for building apps for both desktop and mobile Windows.
    • Only partially true

      WinDiv did this, but the DevDiv never did. Servers and Tools is still all about .NET. If anything, even more so than ever.
  • Commenters Win

    The comments on this page are more informative and intriguing than the actual article.
    Daniel Foerster