Standards drive the market. That’s a truism if ever there was one. But that doesn’t make it a simple truth. There are de facto standards, closed standards, open standards and lots of gray areas in between.
Linux adoption is still growing at a remarkable pace, considering the state of the economy (or perhaps because of it). But the lack of Linux standards, de facto or otherwise, is preventing the operating system from reaching its full potential.
That’s where the Linux Standard Base (LSB), a nonprofit standards organization blessed by Linux creator Linus Torvalds, is supposed to come to the rescue. LSB moved at a glacial pace for years, but it recently picked up speed when Scott McNeil took over the reins.
Now executive director of the Free Standards Group, paid by IBM Corp. as an artist in residence, McNeil is no stranger to Linux, having paid his dues as president of SuSE Linux and in various roles at VA Linux.
McNeil knows that I had pretty much written off LSB as irrelevant, so he recently dragged me to a room full of representatives from companies like Hewlett-Packard Co. and Sun Microsystems Inc. to convince me otherwise. I left both encouraged and disappointed.
Before I explain, let me put to rest the common myth about why LSB needs to exist. Linux isn’t in danger of splintering into incompatible versions the way Unix did. I could explain why the market dynamics that forked Unix in the past no longer apply.
But here’s a simpler answer: The Linux kernel source code supports almost every platform imaginable by default. IBM, Sun and others controlled their own versions of Unix for their hardware. Linux runs on their hardware whether they like it or not.
It’s no secret that Linux distributions like Red Hat Inc. and Caldera Systems Inc. have incompatibilities. But the differences are more like those between Windows NT and Windows 2000 than those between Solaris and HP-UX.
That’s important to understand because people who tend to predict Linux fragmentation usually do so against the backdrop of Microsoft Windows. But the idea that Windows is a consistent standard seems downright silly, even if you’re a Windows fan.
Not only do some applications require specific versions of Windows, some programs won’t even install on the target version of Windows until you have installed the latest service pack.
But although Windows is neither a consistent nor open standard, it’s certainly the de facto standard, at least on the desktop. That’s why LSB needs to exist.
Independent software vendors (ISV) and developers aren’t so stupid as to think Windows will free them from compatibility issues. Windows simply has enough market share to motivate developers and ISVs to deal with the incompatibilities. Linux doesn’t have that luxury.
As a result, ISVs either refuse to develop for Linux or they pick one or two distributions to support until LSB solves the problems.
This is where I get to the good news. After years of delays, LSB has finally produced a 1.0 specification. LSB is also making great strides with internationalization standards. The potential for a global customer base really sweetens the pot for ISVs. Sadly, LSB 1.0 isn’t even close to being as comprehensive a standard as it should be.
But I am impressed by the enthusiastic support by the likes of HP, IBM and Sun to keep things moving. I was especially intrigued to find that HP is intensely interested in helping LSB give Linux the credibility it needs to become the de facto standard server operating system for 64-bit Intel platforms.
I learned this before the HP/Compaq Computer Corp. merger, so it will be interesting to see if this goal was conceived in anticipation of the merger, or if the merger changes it.
Unfortunately, too few people in the room understood that they need to start thinking more like Microsoft Corp. LSB needs to generate confidence by announcing a long-term road map with specifics about things like directory services and desktop standards that ISVs can start planning to support.
Granted, Microsoft can get IT departments to revise their budgets without producing anything more substantial than a press release. LSB doesn’t have that kind of clout. But these Linux supporters gave me the impression they were afraid to create a detailed road map, let alone announce one.
Suck it up, guys. You can always adjust the specifics as needed. But don’t expect anyone to support your plans if you don’t make any.
Petreley, a computer consultant and writer in Hayward, Calif., is the founding editor of VarLinux.org (www. varlinux.org). Reach him at nicholas@petreley.com.