San Francisco?At a somewhat anemic LinuxWorld 2003 conference in San Francisco, big-corporation Linux supporters showed that Linux has grown up and “gone mainstream,” according to Peter Blackmore, Executive Vice President of the Enterprise Systems Group at Hewlett-Packard, who gave the opening keynote Tuesday. And it’s true, Linux has grown up. It’s good enough as a server operating system to have garnered a growing share of the corporate mission-critical server market. The government of China has decreed that all public sector servers and desktops will run Linux. The GNOME project reports that the regional government of
LinuxWorld is driven by enterprise-level software vendors such as IBM, Sun, and Novell, and by budding enterprise open source vendors such as Red Hat, MySQL, SuSE, and others?all of whom are trumpeting the impending triumph of open source over “that other OS vendor,” and all of whom profit at least in part by capitalizing on the fact that Linux is free.
The key word here is “free.” Because it’s free (sort of), there’s a groundswell of adoptions for open source in the public and military sectors that will accelerate as politicians and the public realize that pushing Linux into schools and governmental offices can, temporarily at least, help stretch their budgets. And it’s for this reason that Microsoft needs to stop fighting against open source and begin targeting Linux with its development tools.
Of course, Microsoft has already targeted Unix with its shared-source, non-commercial Rotor project, an ECMA CLI, and ECMA C# language implementation that runs on FreeBSD, Mac OS X 10.2, and Windows. You can download Rotor, look at and experiment with the source, and alter anything you like?as long as you don’t sell or commercialize the result. But FreeBSD and OS X versions aren’t Linux versions, even if they’re close. Shared source isn’t open source, and commercially restricted versions don’t help Microsoft developers deliver software for Linux. So far, Microsoft isn’t targeting Linux publicly. Rotor simply shows that Microsoft could target Linux if it wanted to. I think it should.
Why Target Linux Now?
The time is ripe for Microsoft to target Linux because Microsoft can’t afford to be pushed out of those markets where open source is being rapidly adopted?important institutional markets such as education and government?because that’s where future developers get their training, and where some of the most innovative research and development occurs.
If students graduate with .NET, IIS, SQL Server, and MS Office experience, they’ll push to keep .NET and MS technology in the companies where they eventually go to work. On the other hand, if their educational experience centers around Linux, Apache, Java, and OpenOffice, they’ll be perfectly happy to skip learning Microsoft technologies altogether.
How Free is Free?
There are more than 100 Linux distributions. These are for-profit bundles of the Linux kernel and selected sets of GUI, installation, productivity, management, and utility applications. Most or all of these bundled applications are free as well, meaning you could download them yourself rather than pay a vendor for them. But most people and most corporations won’t. Why not? Because the free versions are unsupported; if you want support, you have to either learn how to fix it yourself, or beg for help from the Linux newsgroups. Neither choice warms the corporate, governmental, or average user’s heart; they want their OS guaranteed and supported by a company, someone they can call (and blame) when problems occur. Home users want a boxed version with a warranty, some directions, and a simple, bulletproof installation.
Jeremy White, CEO of Codeweavers, Inc. gave the analogy that Windows is like a Sears pre-built house. You can buy one and move right in. It’s a fine house, but it costs money, and the design is fixed. Sure, you can alter the paint, carpet, and furnishings, but you’re stuck with the basic plan. In contrast, Linux is more like free lumber. You can build any plan you wish, but you have to build it yourself.
In addition to the free-but-packaged-for-profit software, there are a number of companies providing third-party utilities and applications programs that run on Linux. These aren’t free or even necessarily open source programs; they’re applications that simply target the Linux OS. In other words, these companies are doing with Linux exactly they’ve been doing with Windows for years: piggybacking on the availability of the OS to sell additional services and applications.
If these companies can target Linux and make a profit, so can Microsoft. After all, Microsoft has extensive experience porting complex software to other platforms (think Apple and FreeBSD). More than that, Microsoft has a user base and a developer base that’s second to none.
In addition, the open source community has been hard at work emulating Windows’ features and repackaging them in open source versions. In his LinuxWorld keynote Tuesday, Jonathan Schwartz, Sun’s Executive Vice President, Software Group, showed a slide detailing how Sun’s Mad Hatter software bundle competes feature for feature with software Windows users have come to expect. Almost every major Windows application now has an open source counterpart, and sometimes several.
For those applications where no open source counterpart exists, or for those who absolutely refuse to give up their Windows applications, you can sometimes run the Windows applications directly on Linux via the Why It’s Not an Emulator (WINE) Windows API implementation. Because it’s not an emulator, WINE runs these applications at near native speeds. Currently WINE runs Microsoft Office, Photoshop, and many other Windows applications.
The point is not so much that open source is copycatting Microsoft but rather that open source vendors understand that Linux users, especially the great mass of potential Linux users, aren’t any different from Windows users. They want the same applications, with the same features, the same ease of use, and largely, the same look and feel. As Linux moves beyond the hobbyist and server space into the corporate and home desktop space, there will be an increasing number of Linux users who genuinely don’t care whether their applications are open source, and in fact would probably rather use their familiar Microsoft applications, if they are available, than retrain on unfamiliar and less mature applications.
What’s Holding Linux Up?
Jeremy White said there are three main reasons why users want to switch to a Linux/open source solution: Money, Money, and Cost Savings.
A recent report concludes that, “while much work remains to be done, desktop Linux is now ‘good enough’ for significant classes of users.” If that’s true, why haven’t more organizations and individuals already made the switch?
One reason is that Linux has hidden costs over time that Windows doesn’t have. Another reason is that there are some critical gaps in Linux’s application spectrum. Noted Linux columnist Russell C. Pavlicek gave a presentation entitled “Development Opportunities: What Key Linux Solutions Are Still Needed?” in which he listed the top 10 solutions Linux needs to improve its adoption rate. Although commercial solutions are solving some of these problems, he feels that they present opportunities for open source advocates and entrepreneurs.
In reverse order (as presented), the missing solutions are:
- Migration tools from Eudora/Outlook to Kmail/Evolution. Although some solutions exist, there’s no enterprise level GUI-driven solution for moving Exchange email to an open source counterpart.
- Anti-virus scanning for all protocols. Although users don’t have to be as concerned (presently) with email viruses on Linux, it’s critical for Linux mail servers to protect Windows clients. Users also need solutions that guard against malevolent code delivered via HTTP and FTP.
- Excel macro conversion. Although StarOffice, OpenOffice, and others can read and convert Excel files, they can’t convert or run Excel VBA macros. As many companies have significant and often critical investments in such macros, they won’t migrate until that work can run equivalently on Linux.
- Photoshop. A robust open-source image editing program, GIMP, already has a powerful feature set, but it needs a cleaner, easier interface, with top-tier Photoshop functionality?a GIMP for professional-level users. (Note: Photoshop now runs on WINE.)
- AutoCAD. Pavlicek characterized AutoCAD as “the big dog of the midrange yard.” Both high-end commercial CAD packages and lower-end open source CAD applications exist, but Linux needs a CAD package with a price-to-features ratio equivalent to AutoCAD. AutoCAD itself has so far declined to port its flagship product to Linux.
- Better default fonts in X11 (Linux’s X-windows implementation). The default fonts bother some people to the point that there’s an Xfree86 Font Deuglification tutorial to help fix the problem.
- Mass deployment/enterprise admin tools. Although Linux ships with many administration tools, it needs enterprise-level tools for managing desktops and servers, pushing upgrades, lockdown and refresh, and performing backups.
- Easier connections to Windows shares and Netware file servers. Linux’s connectivity software, SAMBA, is extremely powerful, but needs a better, easier interface. Linux needs a simple Windows domain login (Xandros is available as a commercial solution), and an equivalent to Windows’ network neighborhood.
- Shared calendaring. Pavlicek calls shared calendaring the “opium of the enterprise,” and says some organizations are simply addicted to Exchange’s calendaring features. Linux needs a true, powerful, multi-platform calendaring server.
- An MS Access-like small desktop database. There are many good databases available for Linux, but nothing that provides users with Access’ ease-of-use. Linux needs a lightweight database with a simple GUI interface that database novices can use to design and populate databases.
Although it didn’t make the top 10, Pavlicek says another oft-mentioned corporate roadblock to Linux adoption is the lack of Visual Basic or any language like it.
Step back for a moment and look at this list of missing solutions, and it becomes obvious that Microsoft could fill most of the shortcomings by porting its own software?and its development tools?to Linux.
Microsoft has the ability to delay Linux adoption by not making its software run on Linux, which is the path it has chosen thus far. But it’s time to abandon that tactic. Even if open source software never truly competes feature for feature with Microsoft products, and even if it’s true that there are hidden costs to Linux that make it just as expensive as Windows, Microsoft can only delay, not prevent, the wholesale adoption of Linux by the public sector, where price is the overriding concern.
As Microsoft showed Netscape years ago, free software is hard to beat. But if Microsoft acts now, it can win in the long run?not by beating open source, but by capitalizing on it, just as IBM, Sun, Novell, and other companies are doing now.
For governments and schools, the switch to Linux is inevitable. And to continue to resist it means that these customers will switch not only from Windows to Linux, but from Windows applications to open source alternatives. Give them the real thing, Microsoft. Give them choice. Port the applications and development tools. Turn the millions of Microsoft developers loose on Linux, and let them build the future on both platforms.