If the paper is accepted, the scientist prepares a "camera ready" version according to the specifications laid out by the publisher. Again, on their own time.
The journal staff reviews the camera ready version, and if anything does not fit specification, they request the author fix the manuscript. I assume the staff gets paid something for this service. The scientist, of course, does not.
The journal then publishes the work online, as PDF or HTML. They pay for the infrastructure.
The work is then behind a paywall. Universities and libraries then pay the publisher to have access to the published version of the paper (usually in bulk).
If the scientist wants (or is required to) make the work openly accessible to the public, the publisher charges an open access fee. This fee is paid by the university or grant (it's typically too expensive for scientists to afford on their own).
On top of this, scientists will usually keep a PDF of the paper (either camera ready or a slightly earlier version) on their own personal or university website or on an archiving service like arxiv. They usually give people links to these rather than the publisher's service, because it's hassle free and available to anyone.
@kondziu not a scheme?
@electrocutie @kondziu oh I just got it! (not this toot per se but the pun behind the choice of the name “Racket” for some programming language or environment or something)
@spun_off @electrocutie No, don't encourage them :<
ah @kondziu the good ole public funding, private profit
Scientist sends manuscript to journal. The preparation of the manuscript is paid for by the university and various grants.
Journal sends manuscript to 2-4 other scientists. They review the work and declare whether its fit for publishing. They do this for free (technically usually within working hours at a university).