Checking in on Componentized Linux

Progeny is working towards a release of Componentized Linux 3. Last week, Progeny's Ian Murdock provided a roadmap for the future of CL 3 and announcement that CL was becoming a fully supported Progeny product. Previously, CL was mostly an internal technology for Progeny use, which the company also shared with the community at large as a "skunkworks" project.

Murdock was kind enough to take a few minutes out of his vacation to discuss Progeny's plans for CL, the Linux Core Consortium, the Sarge delay, Ubuntu and other topics.

The company is focusing on the Linux Standard Base 3.0 specification for CL 3 (the CL version number tracks the LSB standard it is based on). A preview of LSB 3 is out now (LSB 3.0preview2), and the final release should be out by the end of Q2, if all goes according to schedule. Progeny is adopting an 18-month release cycle for CL, to track the LSB schedule.

There are a few other changes with CL 3 as well. According to the roadmap, CL 3 adopts a "hierarchical component model," which allows a component to contain packages or other components. This allows developers to build a component from a collection of other components. The new feature will be used "to subdivide the relatively coarse-grained LSB component into a number of finer-grained components" to make the CL 3 release a "better platform for building small-footprint distros for resource-constrained or embedded environments than CL 2.

In addition to technical changes, the company is also looking at a "shift away from services, more towards a product" with CL 3 that would allow customers to create their own custom distributions. Using Progeny's "component compiler," Murdock said it should be possible for a developer to do their own custom distribution "within 20 minutes, 30 minutes". This sounds like a great tool for companies that need a customized distribution, but what about Progeny? If Progeny shifts to the product model, as opposed to direct services, how do they plan to continue to make money? By putting the development tools directly in the hands of their customers, what will they need Progeny for? Murdock said that Progeny would still deliver something of value to its customers.

Murdock said that the company is looking at delivering components "in a form of a service...delivered across some type of authenticated API", which customers would pay for over time -- a sort of subscription service. He noted that the details of this have not been worked out yet, and that Progeny wants to "compete on adding value, not on putting up arbitrary restrictions. We want people to pay us because they're getting value." He also added that if another company could deliver better service than Progeny, "we deserve what we got".

Since Componentized Linux is based on Debian Sarge, which is still unreleased, we asked if the delay had caused any problems for Progeny. Murdock said that the delay "is causing problems for all organizations that depend on Debian, [but] it doesn't affect us more than the others".

After our conversation, Murdock noted on his weblog that Debian "needs to get Sarge out the door as soon as possible, and once Sarge is released, Debian should adopt a time-based release cycle as well. If the GNOME project can do it, there's no reason that Debian can't too".

The company is prepared, no matter what happens with the Sarge release. If Sarge has not been released by June, but the release is "imminent," the CL release may be delayed to wait for the final release. If not, Progeny will base CL Core 3.0 on "a late June snapshot of sarge and incorporate the final Sarge release into a later point release".

We were also curious about the status of the Linux Core Consortium (LCC) project, which has been oddly quiet since its inception. The project was scheduled to release the "common core" during the first quarter of this year, a target that it won't be making, according to Murdock. Part of the problem, of course, stems from the merger of LCC members Mandrake and Conectiva, which has no doubt taken some of the focus off LCC while the companies finish their integration. Murdock said that the LCC is still working towards a release, and that "it actually works out for the better anyway, because we can jump right in to LSB 3.0 without an interim 2.0 release". He also said that the LSB 18-month release cycle "is exactly what we wanted for LCC as well".

According to the roadmap, CL Core 3.0 would include the RPM platform as well as the Debian platform if the LCC development team makes its schedule.

Murdock has also recently made a few comments about the compatibility of Ubuntu packages with Debian Sarge. Murdock says that "A package built on Progeny should work on Linspire; a package built on Linspire should work on Ubuntu; a package built on Ubuntu should work on Progeny." However, Ubuntu packages do not always "just work" on Debian Sarge, which can be a problem given that Ubuntu is gaining in popularity rather quickly.

His suggestion is that Ubuntu, and others presumably, use a compatibility layer to allow packages to work on multiple Debian-based distribution. He notes that he's "a big believer" in what Ubuntu is trying to do, and also said that he's been in talks with Mark Shuttleworth of Canonical about Ubuntu about "Debian-derived distros and compatibility", and also hinted that there may be an "LCC-like" effort for Debian distributions in the next few months.

Developers should be able to get their hands on the first CL 3 preview on or around April 22, according to the roadmap. The preview release will be "essentially the same as CL RC2" but with its packages updated to the current Sarge packages, and with subsequent releases tracking Sarge as it continues towards a final release.