[Panurge]

Actually, some of these problems have already been addressed--or rather, are starting to be addressed. The scientific community has certain means of ascertaining whether or not claimed results (in an article, I mean) are replicable; refereeing can therefore be untethered in some degree from official controls such as are represented by the journals. Things are not so clear-cut in the humanities, unfortunately.

Some libraries, which these days tend to cooperate under the umbrella of the consortium, have joined together to share storage space and costs and arranged to archive print materials jointly. This makes sense: rather than throw everything away or have each library try to save everything, the conjoined institutions identify what needs to be preserved and contribute to a single facility, making sure that a couple or three copies of each book provide some redundancy. Then individual libraries can safely dispose of their own copies in the knowledge that at least one or more will always be available.

JSTOR promises its members a backup copy of everything (in whatever electronic storage medium proves to be current in the future) should the organization, which is not-for-profit, ever cease to exist. An allied organization, called Portico, is now actively archiving electronic materials--especially those journals and other publications that have never been published in print form--and guaranteeing access to them should that become necessary. Individual libraries contribute an annual fee to help fund it (since it, too, is not-for-profit); therefore, it is more or less a form of insurance for libraries' electronic collections. There has been some encouraging progress on this front recently, so there is no need to despair for the future.