Originally Published: Wednesday, 22 March 2000 Author: Matt Michie
Published to: featured_articles/Featured Articles Page: 1/1 - [Printable]

Stick a Fork in It

What ever happened to the Linux Standard Base project? It was originally announced two years ago by Bruce Perens on comp.os.linux.announce. Even after initial backing by every major Linux company and the Linux community, there still isn't an available Linux standard. Does Linux even need one anymore?

   Page 1 of 1  

What ever happened to the Linux Standard Base project? It was originally announced two years ago by Bruce Perens on comp.os.linux.announce. Even after initial backing by every major Linux company and the Linux community, there still isn't an available Linux standard. Does Linux even need one anymore?

1998 was a strange year for Linux standards. The community seemed to have an attitude of Linux inevitability and often of Linux invincibility. Each time an anti-Linux article hit ZDNet, hordes of flames would descend upon the poor author, and Slashdot.org would seethe with proposals to fix any perceived Linux shortcomings. So it was with Linux Standards. The saying was, "First, they ignore you, then they laugh at you, then they fight you, then you win." In '98 the industry rags would laugh at Linux's lack of standards. The Linux community responded with a couple of standards projects. After much flaming, the Linux Standard Base emerged as the leader.

Initially, the community was vocal and active in the decision-making process. There was an excellent series of Freshmeat editorials with dueling opinions. The President of a company called VA Research, Larry Augustin opined, "the most serious threat facing Linux today is Linux." He went on to say that binary compatibility across distributions was crucial to the success of Linux. Two years ago, that seemed true. Most Linux users desperately wanted ports of popular Windows games and applications. It was an uphill effort to convince companies they could make a profit with Linux.

Fast forward a couple years and we have Linux companies with multi-billion dollar market capitalizations. Open source projects like KDE and GNOME have spawned simple to use desktop environments and the promise of office suites. Corel not only ported their office suite, but created their own distribution and added functionality to WINE, the Windows "emulator." No longer is ID Software the only game company with a Linux port. Now, an entire company, Loki Entertainment Software, exists for the sole purpose of porting games to Linux. All this happened without a Linux Standard Base.

The fears of forking never really materialized. As long as the source is available, the minor incompatibilities between distributions can easily be worked with. Some open source advocates could argue that lack of a standard has helped free software at the expense of binary software, which is usually a good thing for Linux as a whole.

The Linux Standard Base is still puttering along. The mailing lists are hosted with Debian and are open for public consumption. However, the LSB seems to be stuck in the Cathedral style of software development. Notables such as Ted Tso and Alan Cox are involved, but there is still no specification available. Even the latest CVS tree seems to be little more than outlines and stubs.

There are some definite benefits to having a standard base to build on. One look at rpmfind.net shows that even among distributions using the Red Hat Package Manager, there are incompatibilities. An RPM built on Red Hat Linux won't necessarily install on SuSE or Mandrake Linux. Although tools exist for converting packages, these are usually used out of desperation. A standard would also make things easier for System Administrators. Even something as "standard" as the X11 configuration files are scattered in different directories on different distributions. There doesn't seem to be a good reason for this.

If the LSB ever finishes their specifications perhaps the vendors will adopt it, but if they take too much longer the market will probably just decide on a de facto standard and run with it. As long as the de facto standard remains open source, this won't cause much harm. However, the community needs to remain vigilant for possible abuses. If LSB wants to remain relevant they need to release a draft specification soon. Ultimately though, the community will decide what is important. So far, a Linux standard hasn't been and doesn't appear to be anytime soon.

Matt Michie is a Computer Science student living in New Mexico. He maintains a small web page at http://web.nmsu.edu/~mmichie.





   Page 1 of 1