The current publishing model is obsolete and broken. To start to fix it, publishing of scholarly/research articles and Books must be treated completely differently. Their only intersection is that the development of both of them CANNOT be based on a scarcity model.
* For scholarly/research articles, the solution is clearcut: Research institutions pay researchers a monthly wage, to do research and produce papers. It's as easy as writing a PDF and uploading it to Arxiv [1]. As a second step, companies like Elsevier, Springer or Macmillan can function as "webs of trust": Getting subscription money for their service, and providing a curation and indexing service as they do now. Shit, they even could provide Editorial/proof-reading services to Universities for people writing the articles. That way, the information itself will be free, and the core value of Elsevier and the others can still be monetized.
* For Books, the "write once get paid forever" model must be stopped. Once the book is written and published, it should be freely shareable. To achieve that, authors should use a model similar to "Kickstarter": Write a TOC, maybe a teaser chapter and look to raise the money he wants to write the full book. (Maybe the book was already written, and chapters are released as their full price is paid). That way the author will benefit in full for the "fair" value of what he wrote, and society will be able to use that knowledge.
[1] I did one: https://scholar.google.com/scholar?cites=1738790492983466495... published only in Arxiv, and it got 9 citations.