October 1, 2003

Warren Togami on the new Fedora Project

- by Joe Barr -
The recent merger of the Red Hat Linux Project with the Fedora Project is an event rife with opportunities for confusion. Although I wrote an article on the Fedora Project just a few weeks ago, I was confused by what I read following the merger. I decided to give Warren Togami (the founder of the Fedora Project and the source for my original story) a little time for all the furor to die down and then turn to him for clarification. So for the past few days, Warren Togami -- with some assistance from Red Hat Desktop Manager Havoc Pennington -- and NewsForge have engaged in both IRC and email conversations to make it all crystal clear.

Let's start by looking at what each project was doing prior to the merger. Togami's Fedora Project provided a single repository with well-tested packages for Red Hat. Using apt-get or yum, Red Hat users could find and then easily install packages not available directly from Red Hat. Mplayer and Wine were two examples. The project team was very keen on testing and had a hard time keeping up with the QA required for submitted packages. Togami was busy with a part-time job and finishing his senior year studying Computer Science at the University of Hawaii in addition to running the project.

RHLP (the Red Hat Linux Project) was less well defined and understood. According to Warren, it was "a vague concept to move the RHL free consumer distribution to become a more Debian-like distribution." Beyond the announcement of the project in July, followed by a retrenchment which saw the shuttering of the RHLP website, not much ever really happened with the RHLP.

But Warren still saw possibilities for the RHLP. He contacted Michael K. Johnson at Red Hat and suggested a merger. He said, "I made the proposal for Fedora to merge with RHLP, because many of the things that Fedora already did for months was within the general ideas of a community based distribution that RHLP wanted to be." Michael Johnson took the idea back to Red Hat for discussion, and Warren began to discuss the possible merger with the Fedora project team.

The result of the merger is the new Fedora Project. With corporate backing and guidance from Red Hat, and a high level of community involvement, Red Hat intends to have two separate distributions: Red Hat Enterprise Linux and a "Debian-style" community-supported distribution from the Fedora Project.

Warren explained it this way: "Enterprise has to do with RH's business side of things, the service and support of the product that customers buy. Free/consumer/community distribution is the same technology, but with no guarantee of support because it is for hobbyists. Companies may choose to use it in production but it is their responsibility to support it themselves, or hire others to help them. This side of the software is more like Debian."

NewsVac editor David Graham asked Warren if he considered the new Fedora to be "a grass-roots or corporate distribution?" Warren replied:

The old fedora.us has been entirely grass roots, with many volunteer developers from around the world that wanted to work on a common problem: easing 3rd party software installation and distribution by creating a single authoritative body.

The new "Fedora Project" we hope will retain and grow grass-roots support, but it is not entirely clear at this moment. There will be a significant amount of expensive capital infrastructure and corporate engineer hours thrown at the project.

This is not to say that corporate == bad. Corporate support makes the combined entity able to handle tougher fundamental problems like server infrastructure (which grass-roots has a hard time paying for) or uninteresting problems like QA testing or implementation of a complex automatic submission & build system.

It is my hope that the combined entity will be a good balance of both. The Fedora Project should be able to harness the volunteer community in order to accelerate the pace of innovation in the hobbyist Fedora Linux distribution, while the corporate side benefits from the rapid pace of improvement in their own business end."

I asked Warren if he was going to lead the new fedora project.
He laughed and said, "I'm overwhelmed trying to graduate college. I wish I had the time." But it appears that Warren and other major players on the old Fedora Project team will continue to contribute to the new project. There is a page on the Fedora Project site with the official word on project leadership: Michael K. Johnson is at the top and there is no mention at all of the original Fedora Project team. But Warren said "Unofficially the most active fedora.us contributors will become part of the advisory committee."

The new project infrastructure is not yet in place and there is much work to be done before it becomes a "going Jessie." In the interim, Warren says that the old Fedora Project will continue to accept packages, put them through their intensive QA, and make them available in the repository.

I asked Warren what the mood was on the old Fedora project team about the merger. He said:

For the most part the old fedora.us team is very excited. The only drawback has been some disappointment from the mostly non-US members about the removal of packages with licensing issues. Otherwise this opens doors of opportunity because a much larger community of packagers will be joining the project during the next few months. The financial resources and engineering experience that RH brings to the to be able to tackle larger necessary problems that volunteers find uninteresting.

Ultimately we will get a lot more done with the newly combined efforts. Community members will have a lot more say in what the software releases contain. This can only be a win for all parties involved.

Asked about his own feelings on the merger, he said, "I personally feel greatly relieved because managing fedora.us, while fun, has been a severe burden on me during these past 10 months. It has been difficult to manage a full course-load at the University of Hawaii, another part-time job, and maintain constant motion in the project. This is a chance for me to release the burden of leadership and devote necessary time to completing my college degree."

As far as turning over the reins of the project he founded, Warren says he has no regrets, adding, "While I will no longer be in charge of the project,I continue to be a contributor in the formation of policies, procedures and standards that will be fundamental to the new project. This greatly reduces my stress because I no longer have to personally initiate many tasks in order to keep the ball rolling. Ultimately I believe the new steering & advisory committee model of leadership will work better in the long run than an over-worked and grumpy dictator."

I wondered if Red Hat had rewarded Warren and the others for their contributions to the project with huge stacks of cash and such. Evidently not, because when I asked if Red Hat had compensated him, Warren said, "
Not at all. I am hoping for at least a hat, but more would be nice. =)(Hint: Not the baseball cap.)"

Joe Barr has been writing about personal computing for 10 years, and about Linux for five. His work has appeared in IBM Personal Systems Journal, LinuxGazette, LinuxWorld, Newsforge, phrack, SecurityFocus, LinuxJournal.com, and VARLinux.org. He is the founder of The Dweebspeak Primer, home of the official newsletter of the Linux Liberation Army, an organization in which he holds the honorary rank of Corporal-for-life.