Very interesting question! I believe that publishing practices will changes dramatically over the next 5-10 years, with open access gaining momentum. This is simply a process that can't be stopped.
Currently there's a over emphasis on impact factor, a measure basically created by librarians in order to decide to which journals they should buy subscriptions. I think this will be replaced by a sort of composite "real impact" or "peer impact" calculated on the basis of downloads, citations, media coverage (incl. social media like Twitter) and more.
Further, since open access leads to a greater impact on peers and media, open access will be an important element in the new impact factors. Since pricing and predatory journals are a potential problem this would lead to reduced pricing and perhaps more non-profit government supported journals.
Never paid for reading, never paid for publishing. The best policy? The only one.
Reading. You would not be saying you don't know places, would you? At least you have a lot of co-authors, collaborators and simply friends all over the world for an intercolleage loan.
Publishing. Journals don't charge. Old ones. Good ones. The new ones do, but rely on people with funds. That's not for us, though it might be easier from the point of view "The esteemed merchant has gold, father. Gold!!!"