Fedora's working groups

Back in September, Matthew posted a call for nominations for a set of "working groups" that would drive the future of Fedora. Each of these working groups would function as a subcommittee of the Fedora Engineering Steering Committee (FESCo) and handle everything that is specific to its particular topic; FESCo's job will become one of handling issues that affect more than one of these working groups:

The current plan is to have five working groups, two of which are directly tied to the "rings" in the original proposal. There will be a "Base Design" group that will manage the core of the Fedora system; this core includes the kernel, low-level plumbing, and the toolchain needed to build it all. The "Environments & Software Stacks" group is charged with managing the next layer, which includes development environments, desktop environments, and more. These two layers provide the substrate that most real-world uses of Fedora will want to have.

From here, it appears that the Fedora distribution is going to fork into three "products" (workstation, server, and cloud), each of which will be run by its own working group. The plans for these groups are intentionally fuzzy until the groups themselves are formed:

Some participants on the Fedora development list were a little disappointed to see that there was not an embedded product on the list. Red Hat's Stephen Gallagher confirmed that an embedded product is not envisioned for now, though, he noted, there will be a place for ARM in the server and cloud variants. Given the current level of Fedora use in the embedded world (or the lack thereof) this position might make sense, but it would be a shame if Fedora were to show a lack of ambition in this area in the long term.

The biggest fuss, however, was certainly about the process by which the groups will be chosen and the degree of freedom they will have. This process was briefly described in the FESCo ticket system and discussed at the October 2 FESCo meeting. In short, FESCo will assign one of its own members to each of the working groups; that delegate will then select nine members from those who have nominated themselves for the position. Nominations are being gathered on this wiki page and will close on October 14, so anybody with an interest in participating should put their name in soon.

How much latitude will the working groups have to define their various products? During the meeting (IRC log), Miloslav Trmač complained about the proposed rule stating that Red Hat employees could not hold more than half of any working group's seats, saying "I'd rather be upfront with that this needs to work for Red Hat". Peter Jones translated that as "basically if whatever they come up with doesn't work for RH, it's a non-starter", and Matthew added "Right. that's a big elephant that underlies everything Fedora does without actually being in our mission docs." That was enough to inspire Jóhann B. Guðmundsson, who has a long history of dissatisfaction with Fedora project governance, to dismiss the entire exercise as "nothing but an utter and total shenanigan on RH behalf". This outburst drew responses from several Red Hat employees, all of whom said that Red Hat had no intention of dictating directions to Fedora. As Matthew put it:

The deliberations of these working groups will be public, so we will be able to see how it actually plays out. Chances are that things will generally go in directions that are agreeable to Red Hat for a simple reason: Red Hat is paying the bulk of the developers who actually get the work done. Nobody could realistically expect that a Fedora working group, no matter how independent, would have the power to direct the activities of Red Hat's staff. So, naturally, directions that are interesting to Red Hat are likely to get more developer time and, thus, to progress more quickly.

In any case, the next step is to name the members of these committees and to let them start the process of deciding what they want to do. The shape of the Fedora 21 release probably will not change much as a result of the new design, but it would be surprising if the releases after that didn't exhibit some new directions. Within a year or two, we could have a Fedora distribution with three distinct flavors, each of which is optimized for a different user community. It will be interesting to watch.