Details changed to protect the guilty. The academic publishing business is undergoing a glacial process of shifting from dead tree products to on-line delivery of journal articles. Nowadays, at least for the journals to which I tend to submit, I as the author end up doing most of the work of writing the text, creating figures and graphs, setting out the layout, editing, spell checking and uploading all the electronic documents directly to the publisher's servers. This saves publishers lots of money and effort since they get near publishable quality (in layout if not research itself) without having to lift a finger.
The journal editors send out the submissions to referees (other authors typically, so I get to referee lots of papers myself) who have to read through each submission carefully and decide whether it is of sufficient quality and interest for the journal in question to publish. Usually a few constructive suggestions are made to improve the readability or point out a few things which could be usefully noted or expanded. Very rarely I referee a paper which needs nothing changed. Sometimes I get papers to referee which aren't worth the electrons/photons used to convey them. There may be a bit of to-ing and fro-ing between referee and author (via the editors who act as glorified messengers) before the editors have a clear cut decision to publish or reject. In all of this, the authors and referees (the same body of people) are contributing their research, effort and expertise for free.
In the end, the journal pays nothing to authors or referees in order for them to do the work for the journal, which then sells off access to the research (and all the layout, figures etc. that the author created) to universities and other academic instituions (who paid the researchers in the first place to do the research). There are some journals which operate on the open access model, where published research is free to download, but the costs of running the journal and online archives is met by the authors. This is somewhat better than the traditional model but still suffers from the problem that authors give up their copyright to their own work, as well as paying the people to whom you gave up your work.
It's not an easy problem to solve, to balance the desire for quality control (I think some form of peer review is necessary, hence the costs associated with running this) and the universal access that the internet provides. The arXiv is an invaluable service for scientists but it does not confer any amount of validation of the worth of the research published. An economic model which respects the rights of scientists to the fruits of their labours yet promotes the advancement of knowledge for all needs to be discovered.
The screenshot at the top of the entry shows you how a supposedly fast-track, electronic only journal handles a submission, they advertise that you get a reply within 33 days of submission. I've refereed for this journal and they expect a referee report within 7 days. It looks as if the editorial board couldn't get their act together and come up with a decision in a timely manner. In the end, luckily they decided to accept the paper so I won't complain too loudly :-).