Looking forward to Fedora 7

Bill Nottingham posted a discussion document on January 4. It keeps the previously-discussed schedule, with the first test release happening on January 30 and general availability of Fedora 7 on April 26. There's a long list of objectives for this release, some of which are:

• Improving the speed of the boot and shutdown processes. "While Xerxes appreciates that he can grab a cup of coffee while waiting for his Fedora system to boot, it becomes annoying when he is not actually thirsty." There are a number of ideas on how this speedup can be effected, none of which appear to involve switching to Upstart. There is talk of replacing init, but nobody appears to own that task currently; it seems unlikely to happen for Fedora 7.

• CodecBuddy - a recognition that not all content can currently be found in free formats. The idea is that the software would detect an attempt to play a file in an unsupported format and respond with an educational session on why free formats are better. Should the user not respond by immediately deleting all MP3 files, CodecBuddy will offer a pointer to available codecs whenever Red Hat Legal allows.

• Encrypted filesystem support, though which encryption technology will be used has not been decided yet.

• Fast user switching - being able to move between different accounts while retaining the current desktop status of each. Making this feature work in a secure and robust way is not trivial.

• The creation of a desktop "spin" of the distribution. That leads to a few related issues - see below.

• Firewire support that actually works. "Requires rewriting the kernel firewire stack. No biggie."

• Support for the KVM virtualization API. KVM appears to be the future of Linux virtualization, so distributions will need to pick it up. What will happen to Xen support is unclear; Xen is unpopular with some of the Fedora folks, but is high on the Red Hat list.

• Support for the new parallel ATA drivers, moving away from the old IDE subsystem. The PATA drivers are an improvement, but they will cause drives to be renamed, leading to potential system chaos. Fedora systems have used the mount-by-label feature for some time, so most installed systems should handle the change without trouble.

• The addition of Nouveau, the reverse-engineered NVidia driver. Whether this driver will be ready by the time Fedora 7 needs it remains to be seen, however.

• Speeding up Yum and RPM. That, alone, should justify the cost of an upgrade to Fedora 7.

There's much more on the list, but the above should be enough to give a sense for what is going on. The Fedora developers would like to improve their distribution in a number of significant ways, and in a very short period of time.

Most of the desired changes are uncontroversial. The creation of a desktop version of the distribution, however, has been the subject of a fair amount of discussion. The Fedora distribution has traditionally been fairly strongly tied to the GNOME desktop. As Fedora tries to expand its community, though, there is a stronger set of voices calling for support of a KDE version of Fedora as well. Nobody seems to oppose that idea, but there is still a shortage of consensus on how it should be done.

As often seems to happen in community discussions, the Fedora developers have gotten hung up on a relatively unimportant issue: naming. Current plans call for the GNOME-based version of the distribution to be named "Fedora Desktop," while the KDE-based version would be "Fedora KDE." The KDE users, who were under the impression that they had a desktop too, think that this naming goes against the idea of KDE being an equal citizen. Others claim that "Fedora Desktop" is meant to be a combination of the "best of breed" desktop software, most of which just happens to come from the GNOME project. They hold out the possibility of a separate "Fedora GNOME" version for GNOME purists; it would feature tools like AbiWord, Gnumeric, and Epiphany, which currently have failed to qualify for the "best of breed" designation. This idea doesn't seem to make the KDE community feel much better.

Jeff Spaleta has posted a call for peace on this issue, saying:

On the technical side, the biggest disagreement would appear to be over whether Firefox should be included. There has also been some discussion of OpenOffice.org and Evolution. In each case, there seems to be some tension between a "pure" KDE system and a desire to include applications that some users are likely to want. Since the unwanted presence (or absence) of any of these tools is relatively easy to correct after installation, one assumes that a solution will be found that everybody is able to live with.

This kind of discussion is not new in the free software community, but it is relatively new to Fedora. As this distribution opens up and accepts more input from outside of Red Hat, there is no doubt that it will get more opinions as well. How these newcomers are accommodated will have a big effect on how successful a more community-oriented Fedora will be. We should see some concrete signs of how well the community is working sometime around late April.