Since 2001 or 2002, I remember a press statement each December: "this is going to be the Linux desktop year". Well, every year that happens to be a mistake. That optimism is lead by a visible improvement of the situation, wider hardware support, more standardization etc...
However, I've realized that that year is not going to happen, never.
Am I crazy? Maybe, but that doesn't have anything to do with this article I'm writing.
Does this mean that I'm suggesting that we should just give up and stop developing any free software desktop environments. Hell, no. Its a gap that should be filled, and if we don't we will fail in showing people what free software platforms are capable to do.
My point is that the model of the desktop is dead itself, and I can't see any big migration from the current predominat home desktop out there. All the application development is moving towards the web, and virtualization as a solution for the lack of flexibility of the current desktop model, is not far from being a reality in corporate scenarios which means that even the desktop is becoming an online service rather than something you may have installed at home. There are some good highlights of this issue on the Brian Setevens' talk at the Red Hat Summit '07.
So, where are we going then? Well, free software platforms are already the strongest option on the most interesting software development market now, the Web. On the desktop side, free software desktop platforms are becoming a serious option on the mobile and embedded landscape, see GMAE, maemo, OLPC, OpenMoko, GreenPhone, and the release of Java as free software. This is, is in my opinion, a much more interesting platform, for the long term, to bring freedom to users than the desktop as we know it nowadays. Free software is now the default choice when no previous choice is made by the broad market.
This is a really promising landscape for the free software world, because it means that Microsoft is not relevant anymore, even if they have a huge market, they are not inventing newer ones. Vista? who cares... That's probably the reason they are talking about IP and patents, Vista is far from being an interesting issue, and it adds 0 or less value over the previous products.
So, as long as the current model gets obsolete, free software and open standards is filling all the new gaps and opening new opportunities. We shouldn't worry about the current desktop market share anymore, what we should do, is worry about providing a good desktop to fill that gap meanwhile we invent new markets where free software can break through and show all its power.
Now the challenge is to reinvent the meaning of the computer desktop concept by developing something new, that solves problems that are not solved already, this is, real innovation.