Presentations, tutorials, etc. by Samba Team members are listed in our:
Back in the last century, the Nineteen Eighties if any other readers can remember that far back, the UK was full of a wide assortment of computer magazines. They became so popular many of them even published weekly editions. I must have purchased and read all of them, feeding my nascent but rapidly growing computer addiction. One of their favorite questions which I remember them asking every few months are so, was “Is this year going to be the year of the LAN ?” (that's “Local Area Network” for you youngsters). Statistics were given about the number of computers sold with and without network cards, and how many it was estimated were connected up to a local file and print server, and if this would grow over the years.
Looking back this has to be the most ridiculous computer question imaginable. Recently I had to go visit the corporate offices of my employer Novell, in Provo, Utah. The hotel I usually stay in has free wireless as part of the service, but not this time. The local network was down. “We've called someone in India” was all the staff could tell me about the estimated fix date. They didn't get it fixed during the week I was there. During the evenings I didn't even bother to turn my Linux laptop on. A computer without a network connection has now become a useless paperweight, not even worth powering up. “Is this the year of the LAN” now makes as much sense as “Is this the year of the telephone”.
All computer magazines are now available over the Internet of course. I can't remember the last time I picked up a print copy (excepting the esteemed journal you're holding in your hands, gentle reader :-). They still love asking questions, the latest flavor of which is “Is this year going to be the year of the Linux Desktop ?”. I predict in ten or so years this will be just as stupid a question as the year of the LAN.
I'm not one of these mad street prophets of the Open Source/Free Software world who will tell you that Microsoft Windows will have disappeared in a magical act of rationality by all its current users. That's about as likely as all of the users in the Matrix waking up at the same time and deciding that being used as a power source for malignant software was a bad management decision. Monopolies that are so entrenched are almost impossible to dislodge, even as they start to eat all the other occupants of their ecosystem.
Look at the current Anti-trust complaints about Microsoft adding competing security features in Windows Vista from their current Security software partners. Nothing has changed from the Netscape or Borland days (there are too many Microsoft ex-competitors to list more than a few here). Microsoft is still using the same methods to remove the competition they used in the last century, poaching their competitors employees and denying access to Windows internal API's. Despite all of this Microsoft's competitors on the Windows platform still find it easier to go out of business than they do to take the radical step of writing software for a different platform.
I think the Linux vendors are going to try their hardest to move into the current Windows desktop space; it's just that it's a really hard place to enter. Without trying to turn this column into an advertisement for my employer, Novell probably has the best Linux desktop offering right now. I know this as part of my job is to help to create technologies within Samba to enable integration with Windows, such as single sign-on with Microsoft's Active Directory. This code goes directly into Novell's Linux desktop product, Active Directory integration is essential for a modern business desktop. Don't worry, I haven't lost my principles: everything I write is Free Software and I fully expect other Linux desktop vendors to pick up and use this code as soon as they do their own QA testing and integration.
But even given the best integration technologies in the world, the business desktop is heavily locked into Microsoft. To give an example, one company I know of (who shall remain nameless) has an essential business application that not only requires Microsoft Internet Explorer to function, but actually requires a (now obsolete) specific version of Microsoft's Java Virtual machine in order to work. Try replacing that with a Linux desktop client without having to rewrite everything. The promise of the open Internet betrayed by thousands of Win32/ActiveX/Microsoft Java developers (all technologies locked into Windows desktops if you're not familiar with the acronyms) who just didn't care about interoperability.
I don't expect Linux to get more than maybe ten percent of the desktops in the developed business world. That number will probably be higher in the developing countries, as World Trade Organization (WTO) rules start forcing places like China to pretend to get serious about unlicensed software. Microsoft are too smart to start forcing the issue however. They know full well the alternative to unlicensed copies of Windows is fully legal copies of Linux, and they'd much prefer the stolen copies of Windows on developing country desktops than their nightmare of a developing country with no possibility of paying their software “dues” to Microsoft.
The reason I think “The year of the Linux Desktop” will be a silly question in ten years is that by that time most of our interaction with computers won't be via the increasingly antiquated desktop metaphor. Embedded devices, starting with mobile phones and spreading out to include cars, televisions, multi-media games, presentations and uses I can't even imagine yet, will become the dominant method of accessing data and information online. Embedded devices with the capability to become information kiosks, phones, “internet terminals” (whatever that will mean in ten years) will reduce the idea of a multifunction business “desktop” to the same quaint idea that everyone would need their own personal secretary. It's a luxury, and one that most businesses won't be able to afford for any but their most expensive employees. The desktop computer simply is too flexible, and does too much for it to be a useful interface device in a world where a disconnected device is an expensive paperweight.
Windows won't become a monopoly in this space. Microsoft's prices are too high, and the overhead of the increasingly bloated Windows operating system will prevent it. Even if Windows can be cut down, Microsoft keep such tight control of it that they won't allow their external programming community the flexibility to change it in the ways it would need to be changed to succeed. Doomed devices like the “Windows Media Center” only go to show how little like a television the desktop computer really is.
But standards are still needed in programming interfaces in order to achieve the economies of scale needed for ubiquitous embedded computing. The only other possibility in this embedded space is Linux. The GNU GPL licensing keeps it from being hijacked by a dominant vendor, as could happen to a similar system such as FreeBSD, and ensures that the base system is a standard platform all can build upon. Once a manufacturer such as Nokia adopts Linux for mobile phones, the number of Linux devices will dwarf any other platform.
I can imagine the guides in the computer history museum in Mountain View showing the crowd of tomorrow around with the introduction “and now we're going to see a computer so primitive it needed to take up your whole desk !”.