we're ~80% of the way through ottawa linux symposium (ols) and the desktop developers conference (ddc) week. thus far it's been moderately useful. the birds of a feather meetings on the second and last day of ddc were particularly good as they allowed people from disparate projects to come together and short circuit months of email conversations with high bandwidth communication. technical achievements have been low, but networking and information transmission has been huge.
today i attended a presentation on perfmon2, a git revision control workshop as well as a presentation on the new dbus interface for bluetooth on linux. all useful for desktop work.
it also occurred to me that a lot of what we're (using the term "we" very broadly) doing in x.org could really use with some attention and participation by the sorts of people who attend ols: kernel developers. we need a preemptive scheduler for access to memory on 3d graphics cards; we need hotplug support for monitors (which ajax has been working on this week and has in some part working already, though i'm not sure poking /proc every few seconds is the nicest long term solution). these are the kinds of things that people working on kernels have been doing since the beginning of time.
i bet if we looked around the desktop we'd see more examples of such stuff. why aren't we reaching out to them more directly and inviting them to bring their expertise and talent to bear more directly on the desktop? historically the desktop has been all about a gui candy coating of software. today it's growing eyes and ears and becoming a lot more integrated with the physical world around it. that may not be the current desktop developers' primary area of expertise, but i'm surrounded by a couple of hundred people for whom it is.