Anything to do with scarcity in terms of copies of any sort of digital media can only come through government-enforced fiat.
It's a unsustainable model... the federal and institutional forces behind maintaining the illusion that copyrighted digital media is somehow logical or even remotely makes sense can only sustain the current popular business models so long.. eventually (one way or another) the market will fight back and bring about prices that reflect the true cost of distributing digital copies (very close to 0).
I don't expect that to happen in our lifetimes, of course. It may take a century or more to reverse current accepted thought in the role of governments to protect irrational markets.
But with software there is still a ton of scarcity.
Talent, knowledge, support, decent documentation, software customizations and so on and so forth. This is how companies like (well there is only one big one so far) like Redhat can sustain making money with software that they give away for free. To put a different perspective on it.. it's not important that software makes money, all that matters is that programmers get proper compensation for their time.
Traditionally only businesses are willing to pay these sorts of costs, however. Which can help explain how 'FLOSS' has penetrated so far into business markets, but have penetrated so poorly into more populist ones. The per-issue cost of business to pay for bug fixes, support, or customization is going to be rather low versus the value of those various items. A performance optimization or a bug fix can save companies thousands of dollars in costs in terms of employee time, material, and downtime.
For individual users, however, this is not so true. Paying a programmer to fix a bug in Gimp or in some other program on a Linux system is just too high... It can cost hundreds of dollars worth of development time to fix a small issue in open source software versus paying 50 dollars for proprietary software.
So the challenge is for open source folks is: In order to create a sustainable model to were people can devote enough time to become experts in the field of open source software development is to figure out a way for the costs of developing software to be distributed over a large number of users in a similar fashion to proprietary software without having to invoke government enforcement as proprietary software does.
Bounties seem a obvious solution, but so far has failed to pan out. When bounties were tried for the Gimp folks and a couple other projects I've heard of they have failed. Maybe it's due to unfamiliarity for the users or the developers being unable to cope with dealing with money when their organization is built around pure semi-charitable voluntary model.
There has to be a market driven solution out there somewhere. But until the open market throws off the shackles of government-driven artificial scarcity I don't think we are going to see it. Open source alone would probably be able to defeat the Microsofts and the Oracles (to the point were they determine that FLOSS is actually sustainable for themselves, not until destruction...) of the world... but it cannot defeat the Microsofts and Oracles of the world working in any reasonable time frame (say a decade or two) with Governments colluding internationally to protect their markets through a bizarre web of laws surrounding copyrights and patents.