BioMed Central (The Open Access Publisher) publishes a large number of open-access journals. The articles are published online as soon as they're accepted for publication in a manuscript-converted-to-PDF format (provisional PDF), updated a few weeks later by the formatted HTML and PDF galleys. Articles are published online only in one volume (each year the journal publishes only one volume; there aren't multiple issues), with articles identified with a number (1, 2, 3, ...) instead of page numbers. A similar trait has been in place with pay-for-everything publishers, something known as Publication Ahead of Print, citations to which are only possible using the DOI (as opposed to the novel system of BioMed which uses article IDs, which are simply numbers from 1 to, say, 26 within each volume, eg. BMC Oral Health 2009, 9:13 (18 May 2009)). In BioMed, an open-access pre-publication history is also available to everyone upon acceptance for publication, containing all the reviews comments and author replies as part of the article data.
AFAIK, such an implementation with the current OJS (version 2.2.3) is a hard thing to do. What I'm suggesting is OJS should be able to have the same behavior, i.e. be able to publish articles not assigned into a traditional concept of an issue, meaning that similar to what BioMed does, a couple of articles are published each month to form a whole volume, no "issue" thing considered. I know that currently with OJS, one has to delete an issue, and then gather a new one with addition of "the new article" and publish again, which I assume changes all of the "publication date" meta-data to the most recent date, rather than the desired "each article has its own publication date."
I know BioMed is BioMed and OJS is OJS, but I still hope the Godfather doesn't prevent OJS from becoming as functional as mainstream industry-standard publishers' scholarly publishing software!