Read other letters about this article
Last I checked, hard disk sizes were still doubling every couple of years or so, and text search and retrieval systems use algorithms and technologies in the public domain. The cost of archiving electronic journals is utterly trivial compared to the cost of archiving dead tree books and magazines, like thousands of university libraries around the world already do. They already manage in the immensely more difficult task of preserving millions of rare paper books and journals; why wouldn't they be able to manage the immensely easier task of keeping a few servers with big drive arrays backed up, powered-up, and running decent indexing software?
And academic publishers are leeches. They do very little actual work; they just shuffle papers between submitters and reviewers, and put the end results on their servers. For this trivial service they charge like wounded bulls, and are free to continue to increase these charges into the never-never because IP laws give them the keys for many decades at a time.