Longhorn's Real Job: Trying to Gore Linux
April 29, 2004
By Steven J. Vaughan-Nichols
SourceYou want to know one reason why Microsoft is taking so long to come out with Longhorn? It wants to make darn sure that it's as Linux and open-source unfriendly as humanly possible. Today, you can mix and match Linux, Windows and open-source programs pretty much as you see fit.
On my home network, for example, I run OpenOffice on my XP Pro, W2K and SuSE, Red Hat and Xandros Linux desktops, while accessing files via Common Internet File System (CIFS) on servers ranging from Windows NT to Server 2003 to a variety of Linux and BSD boxes running Samba 3. I can do that because open standards-based programs enable me to pick the best possible programs for my uses.
While I may disagree with Sun Microsystems president Jonathan Schwartz on many open-source issues, such as his calling Red Hat a "proprietary Linux distribution," he's dead right on one thing: Open standards are what make computing really go.
Microsoft, of course, sees it another way. In the past, it set its operating systems so that unless you ran Microsoft's own programs, you got a second-class experience. We saw that with Netscape.
Now, after having their hands gently slapped by the Department of Justice, the boys from Redmond have another plan: Make it so that users of their next desktop system won't be able to use non-Microsoft-blessed servers or programs at all.
I'm not the only one who sees this coming. The GNOME Foundation and Mozilla Foundation teams also see it coming, and they're trying to come up with a plan of defense.
They're convinced that Avalon, Longhorn's presentation subsystem, and XAML, Avalon's Extensible Application Markup Language, will lock in users to Microsoft's proprietary programs or only to programs written with Microsoft-proprietary tools. Me, I think that's a big deal, but I think WinFS, Longhorn's file system, is at least as big a problem.
You see, Microsoft is busy patenting everything it can lay its hands on with all three. In fact, Microsoft is now building up its patent arsenal, applying for a rather amazing 10 patents a day. The idea isn't to ensure that Microsoft makes a fair profit from its patents; it's to make sure that no one else can write fully compatible software.
The irony of this is twofold: It's exactly the opposite of what the patent system was supposed to do, and XML was meant to open internetwork and interapplication communications, not provide a way to lock them up.
The open-source leaders are considering ways around these problems, such as promoting the use of open standards-based technologies like GNOME's open-source GUI toolkit, GTK+ and Mozilla's User Interface Language, XUL.
Or, taking a leaf from the Samba crew, just try to reverse-engineer and clone Avalon and XAML. Either approach could work, but work needs to start sooner rather than later, no matter which direction developers end up going in.
Next Page: I fear WinFS may be a harder nut to crack.
Miguel de Icaza, one of the most respected open-source programmers and father of the Mono project, which attempts to bring .NET to Linux, says it best. He writes in his blog, "What makes Longhorn dangerous for the viability of Linux on the desktop is that the combination of Microsoft deployment power, XAML, Avalon and .NET is killer."
He's right, but this time, unlike the thumb-fingered proprietary opposition that Microsoft has faced in the past, open-source programmers aren't going to be caught napping.
We already have Mono for .NET compatibility, and I think XAML and Avalon are doable, but I fear WinFS may be a harder nut to crack. File systems are basic, and without compatibility on that level, getting CIFS-based servers to work on a WinFS-dominated network will be a real pain.
Microsoft, you see, is electing to make WinFS not just a mere file system but a complex database engine application that will manage relational and XML data as well as file data.
Microsoft says this aims to give users a way to search for information content independent of format. I say that's a job for search engines, and Google's doing just fine, thank you very much. I don't see any real need to deconstruct something as basic as a file system and replace it with such a complex infrastructure except to make it harder for anyone else—say, the open-source community—to make WinFS-compatible programs and servers.
If this all seems a little far-fetched, well, you tell me, what is Microsoft really trying to do with Longhorn? I know what I want from an operating system, and I don't think I'm that different from most people. I want my operating system to be fast, stable, secure and to work with open standards and have an open set of application programming interfaces (APIs), so I can be sure that I'll have many software program choices. Now, look at Longhorn.
With all of those proprietary innovations, I don't see any way it can be fast, even if we're running Pentium VIs at 4GHz when it finally comes out. Stable? Secure? Please! They didn't get it right in Server 2003, and it didn't have a quarter of the new stuff they're planning for Longhorn. And as for opens standards, come on, this is Microsoft we're talking about after all.
All in all, there's no question in my mind that Longhorn is meant to do two things. One is to make money by replacing aging XP installations (and maybe this time Microsoft will get the Windows '98 users to switch, too!) The second is to make life harder for open-source advocates.
Fortunately, open-source programmers have at least one ace in the hole: Microsoft has never met a product schedule it could meet, so they'll have even more time to get their programs and tools ready to address Microsoft's latest attempt to maintain control over the PC.
eWEEK.com Linux & Open Source Center Editor Steven J. Vaughan-Nichols has been using and writing about operating systems since the late '80s and thinks he may just have learned something about them along the way.