Another possible licensing method is one used by Ghostscript: regressive licensing. The most recent iteration of the product is commercially licensed, but earlier versions of the program are available under GPL (or GPL-like) licensing. This doesn't contradict the terms of the GPL since the transition is from a proprietary-code license to the GPL: each successive revision of the program is at first licensed in a commercial fashion, and then later on released under the GPL when a new revision is published. If a company is developing something where they want to keep a fairly tight rein on the most recent version of the code, this sort of licensing would provide a built-in way to allow things to be gracefully released to the community without compromising initial development efforts.
This last option is the most direct route I can see from proprietary intellectual property to a FOSS model. I don't believe this could be applied to every software product out there that's currently only licensed in a proprietary fashion (e.g., Java), but it might serve as a good first step -- for instance, for newer projects that don't already have a full-blown commitment to FOSS and won't receive as much criticism for switching licensing models.
The open source community needs a bit of reflection. More than anything else right now, the community needs to think long and hard about what it wants to achieve -- and most importantly, what price it is willing to pay to reach those goals.
I'll start with a definition. By "open source community," I mean those who are most vocal about evangelizing and enacting the open source philosophy. They set trends that others follow and wield great influence over the field. They have gone a long way towards making open source software not just viable, but highly desirable as both a goal and an artifact.
One of the things about any given philosophy is that it becomes too tempting to try and view everything through that particular lens. When you have a hammer, everything looks like a nail -- or another hammer. This goes for open source as much as anything else; consequently there needs to be tacit acceptance of the fact that if a group chooses not to release a piece of software as an open source project, that is a fair choice. It doesn't automatically imply a moral failure, and the energy spent attacking them for such things could be better directed into building an alternative.
There are cases where putting on pressure to release a given product as open source may be justified -- for instance, as a way to maintain support for an expired product. But such things should not be approached capriciously, and they also need to be tempered with the understanding that not everything can be released as open source without incurring potentially irresolvable complications. Not long ago I wrote about an effort to get OS/2 open sourced, if only as something to be studied in a legacy fashion. It might be possible to do this with the PowerPC edition of the OS, which has fewer connections with Microsoft's code (one of the big obstacles to opening OS/2), but the amount of work involved is daunting, and the payoff questionable.
The open source community works best when it sees the world as it is now and offers tangible, positive, unmistakable benefits for the open source way of thinking. A good open source product is the best possible argument for open source. To an extent they've already done this with projects like the desktop-oriented versions of Linux, OpenOffice.org, and Firefox. These are all good-to-great things which people can get their hands on and reap immediate benefits. There is no ambiguity about their utility; they speak for themselves, and no demonization of the competition is required to make them work.
None of this should be construed as an excuse for creators of proprietary software to assume that their way is always best, either. Still, there needs to be more of a sense that open source software, while often the better choice, is not the only choice.