Pop quiz: Name the one company that will have the greatest influence on Linux's future on the desktop. Red Hat, thanks to their dominant brand? How about IBM (or any of the other Big Companies Who've Recently Discovered Linux™), with their seemingly bottomless pockets and religious conversion to open source? Or can we cheat a little and say it will really be two companies, thanks to the desktop competition that's shaping up between Caldera and Corel, a.k.a. "Linux's Killer C's?" Or will it be Microsoft, in one form or another, a tantalizing possibility?
Pop quiz: Name the one company that will have the greatest influence on Linux’s future on the desktop. Red Hat, thanks to their dominant brand? How about IBM (or any of the other Big Companies Who’ve Recently Discovered Linux™), with their seemingly bottomless pockets and religious conversion to open source? Or can we cheat a little and say it will really be two companies, thanks to the desktop competition that’s shaping up between Caldera and Corel, a.k.a. “Linux’s Killer C’s?” Or will it be Microsoft, in one form or another, a tantalizing possibility?
In my opinion it’s none of those possible answers, although the last two contain an oblique hint. It’s Inprise (né Borland), now a part of Corel, a fact that requires a quick side trip to Redmond for an explanation.
It’s the Developers, Stupid
Wail all you want about Microsoft’s technical tackiness and questionable business practices, but one fact is inescapable: They’ve understood from the time they left the corporate womb that the way to lock users into your operating system is to get them hooked on platform-unique applications. But since no company can possibly develop all the applications needed to monopolize a market, Microsoft’s biggest allies have been programmers. And that’s why Microsoft has taken such great pains to woo them every way imaginable.
I’m not suggesting that Microsoft is perfect in this regard — they’ve done plenty to make the lives of Windows programmers miserable over the years, and I have the scars to prove it. Between incomplete documentation, platform bugs and incompatibilities, and Microsoft’s “API du jour” approach to system design, being a Windows developer is a remarkable test of one’s conviction that programming really is a better occupation than Olympic javelin catcher.
Microsoft has made the most out of having control of both the platform and its development tools, however. They have showered developers with beta copies of Windows, documentation, and SDKs (system development kits), and flattened as many barriers to entry as possible for developers. They made it easier not just to write Windows programs, but (if you wanted your product to sport a Windows logo) to write programs that relied on Windows-specific functionality, which in turn leads us to the duality of their lock-in strategy: Applications that are fully integrated with the Windows environment make users happier and less inclined to look for an alternative. But those applications are also locked into Windows, since it’s much harder to port them to another platform without a loss of features. Much of this application lock is traceable to the wide-spread use of Microsoft’s programming framework, MFC (Microsoft Framework Classes), a kind of language atop a language that simplifies writing Windows applications.
Back to Borland
What does this have to do with Borland? A lot, because they’re porting their Delphi and C++ Builder development tools to Linux, as part of their Kylix project. Far from being just mere IDEs (integrated development environments), these tools will bring a pair of powerful RAD (rapid application development) environments to Linux. Even more important is the fact that the current users of these packages have long used Borland/Inprise’s own application framework, VCL (visual component library), which will be ported to Linux as part of Kylix. That, in turn, means that tens of thousands of Delphi and C++ Builder programmers will be able to port many of their existing projects to Linux with no more effort than a recompile, giving them a fast, cheap, and painless transition to Linux. If you’re a programmer but don’t happen to be familiar with VCL and Borland’s tools, I suggest you check out the Windows versions of Delphi and C++ Builder at http://www.borland.com, or better yet, get a hands-on demo. Once you see what they’ll bring to Linux you’ll want to repeat one of the better lines from the movie The Matrix: “Buckle your seatbelt, Dorothy, because Kansas is going bye-bye.”
There’s a lot more to good programming for Linux or any other operating system than can be provided automagically by any development environment or application framework. VCL won’t help programmers with many of the operating-system- or desktop-environment-specific features they need to add to their programs to produce the kind of highly usable and useful software we all want. But easier is better, so unless Corel/Inprise/ Borland bobbles this project badly, which I don’t expect, based on my conversations with them and the demo I saw of Kylix, this industry’s original “barbarians at the gate” could literally revolutionize the Linux revolution.
Even that tidal wave of new and ported applications is only part of the story. Borland is part of Corel, of course, which has their own distribution aimed squarely at the mainstream desktop users (namely the people who are currently running Windows and are sick of system hangs, crashes, and forced reboots). They also have a first-rate office suite for Linux, WordPerfect Office, so adding Borland completes the picture. Any two of those three parts would make for a formidable company in the Linux market, but having all three under one roof is far more interesting. If you still doubt the importance of controlling development tools, think back a few months to a deal with similar overtones, the very first acquisition by post-IPO Red Hat. They bought Cygnus, which developed the suite of development tools that Linux itself is based on.
Mini-MS in the Making?
We don’t know yet how effective Corel’s proto-empire will be in exploiting the situation they’ve neatly constructed for themselves, and some people are already shouting from the rooftops that we’re seeing the birth of Microsoft Version 2.0. That’s silly, and it’s far too early to make such judgments unless you really like throwing darts blindfolded. But there are three signs we can look for in the coming year that might tip us off to how this will play out:
First, will Corel’s development tools include just enhancements for working with Linux, or will they also provide features that can be used only with Corel’s version of Linux? Even though Corel Linux is based on Debian and KDE, it has a conspicuously non-standard file manager, and it would hardly be a major surprise to see that user interface drift even further from standard KDE once Corel Linux and KDE both reach version 2.0.
Second, how well and how quickly documented will the differences between Corel’s Linux and the standard components be? Having the first shot at this sort of information was one of Microsoft’s greatest weapons against their competitors over the years. Even though they provided outside developers with considerable help in learning about and using new Windows features, they still had a huge head start on the competition when it came to exploiting those changes in their own products, most notably Office.
Third, if Corel falls prey to Redmondian delusions of grandeur, how does the market respond? We can’t assume that the Linux users will rise up and reject Corel or any other company that gets too assertive. The mainstream market is many times larger than the current Linux customer base, it spends more money per user on software and services than the current Linux users, and most important of all, it’s much more willing to trade off choice for stability and having a “name” and a company behind their product. This is the same market that put up precious little resistance to Microsoft’s assault, after all. The Linux user-base demographics will shift dramatically over the next six to 12 months as waves of Windows users convert to Linux. It’s a virtual certainty that their collective voice (read: spending patterns) will overwhelm that of the traditional Linux user, even if it is not clear yet how much they’ll be influenced by their adoption of Linux.
Bonus predictions: Corel will remain a good corporate citizen, much to their users’ and stockholders’ (eventual) delight; Linux will continue to evolve and improve at a stupefying pace (but it won’t fragment); the Linux user base will go mainstream faster than anyone can imagine right now, with at least some consequences that even I won’t attempt to predict.
Never before have we seen technology, economics, and sociology combine in precisely the way they’re about to in the Linux market. That will make the next year either the most exciting time since computers moved to the desktop or downright frightening, or both, depending on your position in the industry. Whatever the case, the basic lesson that we should learn from Microsoft, that a platform’s success in the mainstream market depends first and foremost on the removal of barriers to entry for users as well as developers, is still true. And that’s why five years from now the Corel/Borland merger will be viewed as one of the turning points for Linux on the desktop.
Lou Grinzo is reviews editor at Linux Magazine. He can be reached at firstname.lastname@example.org.