Longhorn before too long?

Longhorn before too long?

Summary: Charles Cooper explained in a commentary piece last week why Microsoft's upcoming Longhorn operating system is so important for Microsoft to get right. There's a lot of these kinds of articles now, as the technology press in the runup to major software releases serves the same function as the tense music that you heard when David Hasselhoff tried to defuse a bomb in Knight Rider.

SHARE:
TOPICS: Networking
51

Charles Cooper explained in a commentary piece last week why Microsoft's upcoming Longhorn operating system is so important for Microsoft to get right. There's a lot of these kinds of articles now, as the technology press in the runup to major software releases serves the same function as the tense music that you heard when David Hasselhoff tried to defuse a bomb in Knight Rider. Can Microsoft do it? Can Trixie get saved by "the handsome stranger" before the train comes into town and slices her like Parma ham?

Cooper is completely right, though. Longhorn is extremely important to Microsoft, and for reasons related to the structure of software markets in general.

That became apparent to me after I tracked down the research that backed some statements made by Ross Andersen, a computer science professor at Cambridge, at the ACCU conference (the ZDNet article didn't provide a link for it, so here it is). Ross Anderson is a specialist in security engineering, but appears to have an interest in economics--and spikes a lot of his analysis with it.

Software is a market typified by network effects. The more people who use the "network," the more valuable that network becomes. Sticking with the traditional definition of network, Internet protocols are valuable because everyone uses them. Moving into a more abstract definition of networks, Windows is valuable because of compatibility with the wider "Windows product market" all centered around the Windows operating system.

Because of the snowballing value of a product that benefits from network effects, though, there are huge first-mover advantages. Microsoft needs to get Longhorn out sooner rather than later, simply because latecomers have an uphill battle if another vendor has already created the technology around which the market has started to coalesce.

Now, I don't think that Tiger is the Longhorn of PowerPC chips. (That's a topic for a future blog post.) Still, the pressure of a shipping product with some Avalon-like features that has a lot of buzz due to the success of Apple's media efforts and is the only credible alternative to Windows for non-technical users, has to serve as blast of hot air at the feet of sprinting Longhorn developers sequestered in Redmond.

So, there's my scary violin music to add to the tech media's pre-Longhorn symphony.

Topic: Networking

John Carroll

About John Carroll

John Carroll has delivered his opinion on ZDNet since the last millennium. Since May 2008, he is no longer a Microsoft employee. He is currently working at a unified messaging-related startup.

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

Talkback

51 comments
Log in or register to join the discussion
  • Vapour!

    All Vapour!



    Buy Tiger
    Reverend MacFellow
  • Timing

    How important is the timing to MS? Even if they meet their current schedule (questionable), it will be nearly 2 years before we see this product. Will it even be relevant then? Is the timing more important than the quality of the product itself--will we see MS strip more and more functionality out of Longhorn as that deadline looms closer?
    tic swayback
    • It will be relevant

      Which was why I mentioned that Tiger is not the Longhorn of PowerPC chips. Longhorn has LOTS of interesting features, and chief among them (IMO) is a full .NET API. That matters, and I'll be talking more about why in future blogs.

      Microsoft MIGHT strip out more functionality, but it doesn't change the essential fact that you are dealing with a .NET operating system (for all intents and purposes), leading to higher levels of security than would ever be possible on systems that still are essentially native-code based.
      John Carroll
      • Selling that to consumers

        ---chief among them (IMO) is a full .NET API---

        You've stressed this repeatedly, and clearly you feel it will have a strong draw for programmers and people creating software to interact with Windows.

        The question is, how does this translate to the consumer? What is the selling point in having a .Net API that would persuade someone to upgrade their OS (and in most cases, their hardware as well)?
        tic swayback
        • Good question

          And something I plan to talk more about. I think the security ramifications matter a lot, as .NET is more secure than native code.
          John Carroll
          • Doesn't it also rely on people outside of MS?

            If the big advantage of Longhorn is the .Net API's, which are supposedly attractive to people writing programs for Windows, then doesn't that make the success of Longhorn dependent on those programmers? Sounds to me like a fairly big gamble if this is really the case, putting your fate in the hands of people outside of your own control. And how long will it take for those revolutionary programs to come out? If they're not available for a few months after Longhorn actually arrives, will this make Longhorn look like a failure since there will be no immediate reason for anyone to purchase it?

            Then again, I don't think it's as dire as that for MS. Given their entrenched market position, anything they put out will sell, particularly since they probably won't give most OEM's much choice.
            tic swayback
          • Fate

            [i]Sounds to me like a fairly big gamble if this is really the case, putting your fate in the hands of people outside of your own control.[/i]

            Microsoft has ALWAYS had to place its fate in the hands of people outside Microsoft. That applies to any company, but particularly to an operating system / software vendor. Every time they update Office, release a new version of Windows or take a gamble on CRM software, they gamble that people will actually buy the software.

            [i]And how long will it take for those revolutionary programs to come out?[/i]

            A .NET program versus a native application might not look any more revolutionary than a native app versus a Java app. However, one is MUCH better from a security and stability standpoint (or rather, more likely, as programming god might be so good that he makes no programming errors whatsoever in his native program).

            Longhorn apps will be more feature rich, because Microsoft is making it a lot easier to make really advanced user interfaces. That's partly faclitated by XAML, and partly facilitates by a much better and cleaner API in general, making it much easier to integrate those advanced features into your products.

            Is that revolutionary? Call it evolutionary, like moving from native code to Java or .NET.

            [i]If they're not available for a few months after Longhorn actually arrives, will this make Longhorn look like a failure since there will be no immediate reason for anyone to purchase it?[/i]

            Microsoft has released working Longhorn code VERY early. I'm betting there will be lots of Longhorn-oriented apps available, though my prediction at this point is as good as your prediction.

            [i] particularly since they probably won't give most OEM's much choice.[/i]

            Microsoft can cease selling WinXP to OEMs, but they can't prevent those OEMs from selling other OSes. Post-antitrust trial, that's the reality.
            John Carroll
          • Control

            ---Microsoft has ALWAYS had to place its fate in the hands of people outside Microsoft. That applies to any company---

            Sure, that's an obvious point, but in the past, MS has done everything within their power to control the situation as much as possible, hence the antitrust violations. If the success of Longhorn were entirely dependent upon 3rd party developers, MS stockholder should be very worried. But, since their monopoly position guarantees them an automatic sales base, I don't think there's much to worry about.

            ---Is that revolutionary? Call it evolutionary, like moving from native code to Java or .NET.---

            And that's why I don't think consumers will really care much. Give them a new paradigm, let them do something different, or do the same thing in a better way, otherwise, most can't be bothered to change. You are counting on increased stability and security to be the big sell for these programs. Look at what the majority of Windows users already put up with in terms of security and stability (as most are using 98, 2000, XP without later security packs). Clearly these things are not going to be huge motivating factors for people to upgrade. As I've said before, this is an OS that people will get when they buy a new computer. That's going to be its main selling point, it comes with whatever you're planning to buy.

            ---Microsoft has released working Longhorn code VERY early. I'm betting there will be lots of Longhorn-oriented apps available---

            Question--how does all of the features being removed from Longhorn affect these developers? If they were counting on Avalon and Indigo or WinFS, do they have to go back to the starting point again? What if these add-ons are even further delayed?

            ---Microsoft can cease selling WinXP to OEMs, but they can't prevent those OEMs from selling other OSes. Post-antitrust trial, that's the reality.---

            Sure, but the reality is that MS is a highly entrenched monopoly. Consumer demand for their products isn't going to vanish overnight, and any OEM who decides to go against MS' wishes is dooming itself. Sure, you see companies like Dell making a few Linux boxes available, but do you see them doing much of a marketing push on them? It's a token gesture at best, one which I'm sure they got MS' approval for before doing.
            tic swayback
          • Selling that to business

            http://www2.cio.com/poll/results.cfm?id=555
            IT_User
          • LOVE that link dude

            However, John will be quick to point out why they must be "deluded" or .NOT getting it (heehee)..
            Jeff Spicoli
          • You posted for our amusement.

            I have to assume you did, because you're far too knowledgeable to use a web poll as proof of anything.

            I'm curious about how the people who answered understood the question. The incorporation of the .Net framework proceeds apace, so I'm not sure what the question means.
            Anton Philidor
          • Note the source of this particular poll

            It is CIO Magazine, and the readership are professionals, and as far as I know vendor agnostic.

            The one thing I can tell from personal observation is that .net is not permitted in the major enterprises whose Technical Reference Models I have reviewed (only six or seven, but they have been consistent on this score).

            This is not a particular bias on the part of big business, it's a matter of practicality. It costs money (staff, logistics, training) to support multiple environments, and management is taking a hard look at IT investment dollars. The days of the open checkbook are past.

            The CIO poll only corroborates personal observation. I'm curious about Mr. Carroll and his extolling this particular technology.
            IT_User
      • .WHAT?

        A full.Net API will give you the levels of security that you should
        have had to start with. The fact you've put up with the lack of it,
        is simply bad judgement. Another two years of poor security and
        then the promise of something better from the company
        responsible for the status quo. Is this good enough for you?

        .Net. Another retrofit and layer of abstraction addd to an already
        bloated system. Why are the system requirements so high for
        Longhorn? My bet is .Net and the performance hit that the new
        run-time will cause. What's in it for the user? Will Word run
        faster as a .Net app? Will I now be able to buy things on-line?

        While business would be increasingly better served by a thin
        client architecture, Microsoft proceeds to thicken it's client.
        Adobe's acquisition of Macromedia puts Flash and postscript
        squarely in the path of Microsoft, and offers solutions that are
        not platform centric. What exactly can .net offer the user that is
        not available now and why should Windows users accept another
        picket in the license fence that surrounds them?

        regards
        Harry Bardal
      • .net is only

        Going to be relevant to Microsoft only programmers. There are other platforms and programs / programmers out there in the world, more than Microsoft I would venture. For the end user at home, .net means nothing. It has no value to them. They don't care. Then end user wants a secure platform that is easy to use and not invasive. They want a platform that is inexpensive as well.

        I just converted 5 families to Linux John. These were die hard Windows people that have had it with all the promises and not ever getting what they expected. I suggested the Mac to these people and they opted for Linux. Why John? Why they wouldn't take the Mac, money. Times are tight for a lot of the average Americans out there. A lot of people are starting to unplug and get rid of their PC's because of the cost of maintaining them. The hassle of using them and the fact they are p!ssed about being hood winked.

        Simple, because they didn't have to buy a lot of new hardware if any at all. All it cost them was my fee for installing and training. These people are now using Linux and actually enjoying using their PC's again. How did they learn about Linux? Simple they saw it on my laptop, asked what version of Windows it was and when I told them it wasn't they wanted to know more. I told them the pit falls of Linux and the benefits. 5 more families made the decision to move from Microsoft to Linux.

        When I asked these average users what they thought (if they even knew about Longhorn) all but 1 said they had heard about it and some of what it is supposed to be able to do. And the four that responded, had this to say: no matter what the PR folks say, no matter what the ads say, there is no way in hell they would purchase Longhorn, and if the PC they had died they would look for an alternative or bite the bullet and go Macintosh.

        But back to my original point, .net is not the be all end all for programming. Not unless it will perform 100% the same without mysterious errors on ALL platforms and architectures out there. And quite frankly based on Microsoft's past performance, I highly doubt that will be the case.
        Linux User 147560
        • You hit the nail perfecto

          John seems to be blindly looking at this from a programmer's perspective. He's drooling over the .NET environment, not even realizing that 99% of the Joe Sixpacks out there don't even know what the hell .NET is. And considering that most all of these people never update Windows or their AV and don't even have a firewall or know what it is, to think they will be jumping for Shlonghorn for the "added security" of .NET just ain't gonna happen.
          Plus, even if .NET does catch on (which it doesn't seem like it's going to), it will take YEARS before people will start to see its supposed benefits. And this is also assuming that everyone stays with Windows so they even [i]can[/i] get something out of .NET. And is everyone going to wait all these years for this to happen??? Shlonghorn is two years away, and it will take a few years for Shlonghorn .NET apps. That's close to 5 years!!! People will be flying their cars by then! What Dr. Carroll doesn't see is that no one else is enamored with Microsoft anymore. Even Bit Byte seems to be starting to turn away, a big testament considering his usual Micro-zealotry. John has the attitude of someone from 1995..
          Jeff Spicoli
        • Not for business

          http://www2.cio.com/poll/results.cfm?id=555

          I have seen the Technical Reference Models for several large organizations, and none have .net in the "move to" catgory. Some have it in "hold," others in "divest."

          I attended a Technical Design Review last year, where the contractor presented a program plan based on .net implementation. The excuse was that the majority of the staff were Visual Basic programmers and retraining or hiring new would incur excruciating delays. The contractor no longer has the contract.

          Bottom line - the cost of maintaining multiple development environments is not acceptable to management. The days of an open checkbook are gone.
          IT_User
      • So what you're suggesting

        is that legacy win32 code must be phased out for new .NET code. In other words, you want MS to give up their #1 (by FAR) advantage over other operating systems.

        While that may be the best thing in the long run, what's to stop programmers from switching platforms altogether? This is why it is important that MS not only get Longhorn out quickly, but produce all its intended features quickly as well. Because if there is nothing to distinguish Longhorn over other platforms (and trust me, the GUI improvements I've seen on screenshots are utterly unimpressive), then there's nothing to stop programmers and users from switching. Especially if XP support is discontinued before all of Longhorn's intended features are ready (although I think even MS would understand that point, even if it does drive them nuts).
        Michael Kelly
        • a small requirement indeed!

          Just rip out yer entire Win32 architecture, ya know, over the course of a weekend or two..and Microsoft thought .NET was going to sweep over everything???
          Jeff Spicoli
      • Too quick to promote....

        tiger may support the .net protocols but it is NOT a .net OS. Longhorn maybe built on top of .net to various degrees, and it may be called a .net os.

        The reason that longhorn is SO FAR away from being a mature and usable product, is IMHO, the complexities of porting the spaghetti nt kernel to a pure message passign / RPC model such as that of .net. Although the xp kernel does use lrpc I doubt if anyone ever designedf it as a whole system, and things simply accumulated on top of each other. Now they are trying to basically re-design the whole thing in a more sensible way.

        Nevertheless, tiger is NOT .net, but apparently supports it.

        -m
        michael-t
  • John, that's exaclly what created WindowsME

    Let's jsut say it was not a shining moment for Redmond.
    No_Ax_to_Grind