I’m sure I’m not the first to propose this, but git is a fine mechanism for publishing academic content. It bundles up a collection of files, gives the bundle a cryptographic hash which can be cited and used to verify the files, and provides a means of easily replicating and distributing them. And it is a free and open standard.
Moreover, GitHub shows how a business can thrive by providing git-related services. They host public content for free, and charge for hosting private content.
An open-access academic publisher could use exactly the same model. Researchers need to publish their work: they need public repositories. However, while they are doing the original research they also need a place to store results and collaborate. This could be done with private repositories.
What makes this relevant today is this (paywalled) paper in Science that describes how a fake paper was submitted to many open-access journals and accepted by a large proportion of them. The implication is that many open-access journals will publish anything as long as they are paid.
I’m not going to go into details about the irony and hypocrisy of a for-profit publication accusing others of milking authors via the publication process; if you want that read Mike Eisen’s take.
What I will say is that I don’t agree with how many open-access journals fund themselves via author publication fees. For example, Eisen’s PLOS journals have publication fees in the thousands of dollars per article (fees can be waived for authors who cannot afford them). The problem is that the actual costs of hosting the publication files should be much lower than that.
The GitHub model shows that there is a way of providing open-access without publication fees at all. Of course, it still relies on authors to foot the bills, but they are getting valuable additional services for their money (private repositories and collaboration tools), and they can skip fees entirely if they so choose.
Notably, and in contrast to the PLOS approach, the git approach encourages replication. Anyone can copy a publication, and the hash ensures the integrity of all copies. This is the only way to have archival publications.