A Look at the UserLinux Proposal

Why would we need another Linux system? It's not as if there's any lack of distributions. We spoke to Perens to get the details on UserLinux. According to him, there's a need for UserLinux because current Linux vendors are too focused on profits, and the needs of users are being neglected. In the last year or so, he noticed "the rise of proprietary open source. Software that is purportedly open source-licensed, but the user is still made to pay in the long run."

This trend is causing problems for Linux. In his white paper, Perens writes:

In creating a community-driven solution for business, UserLinux could provide an alternative for businesses that aren't looking to pay per-seat fees to companies like Red Hat. However, a community-driven project will face problems that Red Hat and other vendors have already (at least to some extent) overcome.

One major obstacle that UserLinux will face is garnering the support of independent system vendors, such as Oracle. Any distribution aimed at the "enterprise" market will need that support. Theodore Ts'o noted that this can be an expensive undertaking for some of the highly desirable ISVs. Perens acknowledges this in the second draft of the UserLinux paper:

The question, of course, is how the organization will pay to support a lab and other endeavors. Perens explained that he would like to see an organization for UserLinux service providers, which could certify providers and serve as a point of contact for the providers and customers. Providers would probably pay some fee for certification "on a sliding scale based on the size of the business" to allow for sole proprietorships. The organization might charge some percentage for business referred to the providers, but he doesn't want the organization to be a mandatory gatekeeper between customers and providers. He also noted that UserLinux could also have uncertified providers, they would simply not be allowed to use the trademark for certification.

The service organization also answers another question that businesses are likely to have: "Who is going to support me?" Perens states in his paper that a organization built around UserLinux would actually be able to support more customers than Red Hat:

This organization may work to provide technical support, though it bears a strong resemblance to Red Hat's early "Support Partner" program, which was never very successful. What may prove harder is convincing potential users that other sorts of support - such as security updates - will be available for several years into the future.

Another obstacle faced by UserLinux is package support. Specifically, which packages to bundle into the distribution. At this point, Linux has several areas where there are a number of competing packages that perform the same general functions. Perens argues that an enterprise project should make choices between various packages and pick a single package rather than be bogged down trying to support a array of packages. This would not prevent users or vendors from adding packages, but the default system would include only one of a given type of package.

This is likely to cause some heated and unpleasant debates as UserLinux moves forward. There are already some strong objections to Perens' preliminary choices on the UserLinux discuss list. Eventually, these choices have to be made, however. Perens proposes that these and other technical choices be made by a meritocracy, similar to projects like the Linux kernel, the Apache project or the Debian project itself.

UserLinux is, at this point, only a concept. There is much work to be done, and much of it is in uncharted territory. Whether or not it succeeds depends on a number of factors, some of which are obvious now and others will only become apparent with time. But if the success of Linux has taught us anything thus far, it is that the open source community can succeed where many expect failure.