One of the ideas that has had currency in Canonical (perhaps in other parts of the Ubuntu community too, I’m not sure) is that getting more applications available to Ubuntu users will be a critical part of getting Ubuntu into mass use. This was well presented by mpt at UDS-N in this presentation (starts about 2:49).
Personally, I think this is the wrong focus. I think it’s critical that things people use every day are brilliant and reliable. My daughter’s number one issue on Kubuntu (our Ubuntu flavor of choice) is random X (actually Intel driver related) crashes. Until the core parts of Ubuntu (platform and high use applications) are rock solid and interoperable with their proprietary counterparts (yes LibreOffice/OOo, I’m looking at you), 23 different solitaire apps, of which 21 are “not very good”, just don’t matter.
So I was pleased to see today a new study that seems to support this view.