On Tue, 2008-10-21 at 13:30 -0700, Thomas Bushnell BSG wrote: > On Tue, 2008-10-21 at 14:59 -0500, William Pitcock wrote: > > If we waited for a release to be 100% perfect, it will likely take > > several more years. The good news is that the amount of inline firmware > > in the kernel is decreasing. So, eventually, all non-DFSG > > redistributable firmware can belong in firmware-nonfree. > > Do we have an ironclad commitment to not add any additional non-DFSG > firmware, period, no matter what? I would accept a compromise which > guaranteed an increasing slope. But not a back-and-forth thing. Your > reply focuses on regression issues, so is that really sufficient? We > guarantee that, say, there will always be *less* non-DFSG firmware in > each release, and we guarantee that there will never be *new* non-DFSG > firmware. Unfortunately, those who contribute to Debian must be dedicated to ensuring future releases of Debian support the latest available hardware at time of release. If those drivers require firmware, then we can work to ensure that they use the kernel's firmware loading framework. This is a cause that is a good idea for many practical reasons excluding ensuring the firmware is segregated to firmware-nonfree. However, not supporting the latest 3ware RAID card due to non-free firmware, as an example, would be unacceptable, considering Debian's strong foundation in being run on servers. Likewise, failing to support the latest 3D hardware or audio hardware when DFSG-free drivers are available, but depend on non-DFSG firmware, will lose Debian users on desktop hardware as well. That said, Debian release cycles are fairly long, so there's time to make sure things are implemented right in the future. > > > If the NMU involves removing support for hardware, then no, the NMU's > > solution would be in my opinion unacceptable, and hopefully enough > > people agree that it would be rejected. > > Thought so. So the claim that "nobody is standing in the way" was > simply false. People are standing in the way of a simple fix for a > simple bug, and insisting on a more complex fix. I agree completely > that the more complex fix is better, but it is simply not true that > nobody is standing in the way of a fix. Rather, they have declared that > only one sort of fix is tolerable, and mostly refused to discuss the > question. People are not standing in your way as long as it does not cause regressions or break support for current hardware that people may wish to use. William
Attachment:
signature.asc
Description: This is a digitally signed message part