I mentioned the idea of an Elsevier boycott here last year. Here's someone who's thought about another course of action: if you can't beat 'em, join 'em - up to a point:
How about this for a solution:
1. We start a (say) monthly journal.
2. Set up a simple, lucrative payment for accepted articles. As a first stab, how about $10,000 paid personally to the researchers responsible. Similarly, each reviewer receives $2,000 for each article reviewed.
I imagine this would be enough to attract some serious submissions. So serious, in fact, that universities and libraries would be obliged to subscribe (at drastically reduced prices). Oh yeah, and you can only submit articles if you or your institution is a subscriber.
So every month hundreds of thousands of dollars would be distributed to scientists, and there would be a monthly prize pool for excellent research. And I bet the journal would be a cracking read (or as cracking as stiff scientific prose ever is, anyway).
But why not take it a step further.
3. Assuming, of course, you executed that first step and attracted some worthwhile research, you could simply distribute all of the profits. So if you attracted 5% of the market… say 50% penetration at half price, you would be able to distribute 250m a year (I know, I know).
I imagine a star researcher would prefer to get paid $200k rather than submit to Nature. And as the quality of the new journal improved, it could even end up becoming more prestigious.
The first problem I have with this, and it's a big one, is that if you're giving "drastically reduced" subscription prices to libraries and universities, where are all the profits coming from that you're using to pay the authors? I believe that squeezing institutional subscribers is, in fact, the business model for many of the scientific publishers. A second difficulty (there are others, too) is that our current system already encourages people to divide publications up into smaller units and get a longer list of papers to their name. If there's sill more incentive to publish (cash rewards!), I can imagine this problem only getting worse.
The general rule in nonscientific publishing, that "money flows towards the author", is based on a market that pays for the finished products voluntarily, and has the option of not reaching into its collective pocket at all. Publishers are also free to pick and choose among the many, many manuscript opportunities they're offered, trying to find those that they think will provide a return on their own investments of time and money.
Scientific publishing is a different undertaking, with a mixed set of motivations. We're publishing our results to show what we've accomplished, and to add to the storehouse of human knowledge (as opposed, say, to adding to the storehouse of human romance novels or slow-cooker chicken recipe compendia). We're also publishing to make ourselves look better, to our current employers and to potential future ones, not that such publications are a very efficient way to do that, but still. And the readers are keeping up with the published matter partly out of a desire to learn about what we have to say, and partly out of professional necessity.
Here's the thing: if it were not for the expense necessary to produce the journals, there would be no market involved at all. In the early days of science, results were distributed by personal correspondence, and journals are, in a way, just collections of such letters. Some of them, you'll notice, refer to that function by having "Letters" in their names. No one was paid to write those letters to their colleagues, and no one paid to receive them. The only expenses were the writing materials, the time it took for composition, and postage - just as the expenses now are the time and effort for composition (on the author's part) and the time and effort for editing, reviewing, and publishing (on the journal's part). These expenses have, in theory, been going down as fewer journals are published on pressed tree sheets and hauled around by mail trucks, but they cannot go to zero as long as there is editing involved, as well as commentary on the papers themselves and publicity for them, not to mention things like server maintenance, etc.
All the fights about the cost of scientific journals are fought on that narrow strip of ground. How much does it cost to do these things, and is it truly worth what people pay for them? Otherwise, we might all just as well upload everything into a central electronic archive, with a nominal fee to keep the hardware running. We'd then be back to sending letters around to let everyone else know when we've put in something that we think is hot stuff, or setting up web sites to get the news out. And some of those might become useful enough to start charging a subscription fee, and then. . .