>Yes, but the last 20 years consisted primarily of a closed source culture.
>When the burden of maintenance is placed on a relatively small group of
>developers, the cost is prohibitive.  Not so in open source where the
>developer and consumer can be one individual.  In this regard, I think the
>next 20 years will be much more interesting.

With all due respect, this view is somewhat naive, and it ignores 
some important history.  (1) The culture in question derives 
primarily from precisely the group of people who gave us Squeak, and 
its a pretty damned useful culture; but more important (2) it arose 
from a land of user interface chaos, which hurt both developer and 
user.

So much for the theoretical arguments.  Customers and users don't 
really want this stuff so much either.  Substantial evidence in 
modern times confirms that markets (this was all about "looks matter" 
and marketing, right?) reject in large part radical departures from 
UI's, and except perhaps for the occasional backdrop image and mouse 
selections, spend very little energy modifying their "environment." 
Institutional clients, in particular, find substantive 
individualization a massive headache to maintain, and customers 
generally don't really want it for their benefit.  Except for the 
"Kai's fill-in-the-blank" products, which are by their nature 
graphically oriented, most roll-your-own interface products in recent 
times have failed.

While the occasional hacker (including myself), likes to mess around 
with this stuff, it is rare that a novel UI is accepted or even 
acceptable.  As noted, both Java and Tcl/Tk eventually abandoned 
their hopes of a unilaterally imposed unified GUI for local UI's.

And so far as "open source culture" being the harbinger of good 
things yet to come in GUI, all evidence is to the contrary.

If there is one area that has stifled and kept OSS out of the 
mainstream, it is precisely the inability to produce a meaningful, 
consistent and usable GUI environment.  The present morass of 
mediocre *nix GUI environments, and the inability to get past petty 
politics to fix the horror of all that variety has been, IMHO, a 
major impediment to the open source movement.