Take copyright away so everyone, including those who currently follow the law, is free to copy as much as they want from wherever they want, and channels like BitTorrent rapidly become more viable. For one thing, more people would soon learn to use them, and their effectiveness increases with the networking effect. For another thing, you don't have to hide your identity if you're sharing ripped stuff any more so the "poisoning the well" tactic by copyright holders becomes almost worthless.
I suspect the optimum model in the absence of any copyright at all may be as simple as free distribution and inviting donations, but given that people have tried this and a tiny fraction of users ever donate anything at all, that's a pretty unappealing prospect if you're the guy looking at investing years and millions in building good software for others to enjoy. I have seen a few other funding/incentive models advocated in similar debates in the past, but usually the administrative burden and risks when you think them through in detail make today's copyright mess look the epitome of elegance.
There is nothing fundamentally uncrackable about the DRM of any games available through Steam. Indeed, many games on Steam are available in cracked versions. Nevertheless, this doesn't change the value of Steam or the ability of Steam to do amazing business. Often times it's just plain easier and better to buy a game through Steam than to go to the trouble of pirating it. That's very much a part of Steam's business model. Embracing digital distribution and adding value to digital ownership, in contrast to the typical approach which is to poison digital ownership in the hopes of crippling it enough to make it conform to the characteristics of physical ownership (i.e. constrained scarcity).
As to open donation models, I think you have a misconception about the idea. It is utterly irrelevant if the number of people who donate anything is vastly exceeded by the number of people who play a game (or listen to music, or watch a movie) without paying anything. What matters is the total revenue. If you can make more revenue by having 2,5, or 10 times as many people donate money even though 100x as many people donate nothing. This situation may tickle your "unfairness" sense (but to that I'd ask, do lending libraries do so as well?) but at the end of the day the artist still nevertheless has paid the rent and put food on the table, and that's what's important.
As far as donations, I do understand that the point isn't whether everyone pays. However, you still need whatever total you raise from whoever does donate to be worthwhile. Given that (a) there is nothing in today's law stopping someone from adopting that approach if it truly provides a better incentive to share their work, (b) various people have tried, and (c) I'm still waiting to hear about the market-changing success stories, I think a major cultural shift would be required before a purely donation-based model could provide incentives as effective as even the broken copyright mechanisms we have today.
The appropriate response is to embrace it as Valve has with Steam. The low incremental costs of distribution allow new business models such as "steam sales". This model has generated literally billions of dollars for Valve and other game makers. And in several instances small game companies have come back from the brink of death.
You can fight against the grain or you can work with it (steam, amazon mp3, humble indie bundle, etc.) The idea that taking advantage of the characteristics of digital goods is akin to surrendering to piracy and thus an end to all revenue for creators is profoundly mistaken. Above all else that's the point I'm trying to make.