I was mostly talking about keeping this in mind when developing new software. It definitely will be hard to make big departures from what's already in existence. It's not impossible though. You can always give users the option to add X toolbar to a window and choose which buttons appear by default and such.
Of course, these solutions don't lend themselves to all software and that's okay. What's really important is just to keep these ideas in mind and not write off design as just unnecessary ornamentation in the future. Obviously we have to be thoughtful about where these ideas would work best. The easiest way to do it is know your target users and design for the least common denominator among them. Your software can launch with a design meant for the least common denominator by default with options which can be displayed and remembered later for more advanced users. I can't repeat this enough though: this can only work on a case by case basis and by knowing your users. Remember too that incremental departures are swallowed much easier.
Design may not be the number one thing holding back FOSS adoption but it's high on the list I think. When the FOSS community can offer equal or better alternatives to what's out there, market share will skyrocket. I think Linux is pretty much there already. It has a few blemishes but the big thing is probably that normal folks are just scared to try it, won't go through the trouble of installing another OS over the one that came with their PC and the appalling lack of computers that come with Linux pre installed.