Seth: So you are looking for more religion in GNOME. Well I think part of the trouble to the degree that there is a real problem (not sure there is) is that people who come up with really new ideas has a tendency to not stick around to complete them, or actually they tend to stick around, but they move their attention to their next grand vision before completing the first. Examples of such in my mind are medusa, dashboard and GPF (gpf in gnome cvs).
Even in the cases where the originator sticks around or a strong enough development team evolves to keep the project going so do innovation also tend to demand a lot of trial and error which has a tendency to demotivate people. I am probably biased, but I think that GStreamer deserves the label innovative. That said it has taken an enourmous effort with each subcomponent being rewritten numerous times, some still need an iteration or two before getting there. Keeping developers happy during these phases have not been easy, which is why we have had some people abandoning ship by implementing alternative backends to their applications etc.
As for grand visions, well I think the HIG do qualify as a grand vision altough maybe not innovative in the sense that HIG’s is an old Apple concept.
I also think Havoc makes a good point in regards to the kernel. There needs to be something to build upon before grand vision stage is plausible. An example would be that for someone to pursue a dream of building an multimedia centric network media system where every machine on the network shares things like music playlists, you exchange files and collaborate using video confercing with built in whiteboards etc., and on and on. Well that vision gets a bit overboard to get anyone else involved in if the system doesn’t even have a basic music playback system.