This is one aspect of what is covered by the term "publication"; it is convenient here to consider it separately from the other aspect, which is discussed under "dissemination".
Libraries have established effective means of archiving books and journals published on paper, which can endure and remain legible for hundreds of years. The archiving of electronic publications is a very different matter, and one for which no solution has yet been established. Storage media evolve rapidly, as do computer hardware and operating systems, and software for reading the documents. I would be out of luck if I needed to access information stored on the 8-inch floppy disks with which our group began our microcomputer work a mere fifteen years ago, and it can confidently be predicted that today's CD-ROMs will not be generally legible fifteen years from now. Before we can in good conscience commit our research results to paperless publication, there will have to be established a system for transferring the documents from obsolescent to current storage media, together with software for reading them.