A new interesting format of open access electronic proceedings called “Electronic Proceedings in Theoretical Computer Science (EPTCS)” has recently been announced (and mentioned by Luca and Suresh):
Electronic Proceedings in Theoretical Computer Science (EPTCS) is an international refereed open access venue for the rapid electronic publication of the proceedings of workshops and conferences, and of festschrifts, etc, in the general area of theoretical computer science, broadly construed.
I admire the very efficient and streamlined mechanism:
We do not charge authors or event organisers for electronic publication in EPTCS in any way. If hard-copies of proceedings are desired, event organisers have the choice of organising the printing themselves or taking advantage of a standard contract we will make with a printing house. Copyright on all papers is retained by the author, and full-text electronic access to all papers is freely available, without any need for registration or subscription. Permanent archival of the journal content is ensured by organising the journal as an overlay of the Computing Research Repository (CoRR): see arXiv.org. In addition, a printed version of the entire series is available to subscribers.
Much flexibility is given on how to use this format with just minimal measures against abuse:
Event organisers may apply for publication of proceedings or festschrifts as an issue of EPTCS by filling in this form and writing to a member of our editorial board whose scientific interest are closest to the topic of that event. Such an application needs to contain a detailed description of the way in which the papers that are considered for publication in that issue are being refereed. The application will be granted only if we trust the event to select scientific papers of quality only. This trust may be instilled through the organisers of the event, the program committee members, and (if applicable) the quality of earlier editions. Once an application has been granted, the refereeing is left entirely in the hands of the committee established for this purpose by the event organisers, and our staff will perform a quick check of format and contents only.
What I like most about this new format is the subversive way that it blurs the meaning of a “publication”. Are papers in these “open proceedings” considered a conference publication? A journal publication? On one hand there doesn’t seem to be any reason to view, say, STOC papers any differently whether STOC decides to use this format or not. On the other hand, since no copy-right transfer is involved and the papers are just arXiv papers, is there any reason to treat them differently from other arXiv papers? On the third hand, maybe EPTCS is itself a journal with issues, an editorial board, etc. E.g., should there be any problem in publishing these papers as is in journals (even economic journals that now disallow publication of CS conference papers)? If I organize a small workshop on my favorite tiny research area, then papers presented there are expected to later (or previously) be sent to established CS conferences; does this change if I decide to organize the papers of my workshop’s speakers into these virtual proceedings?
I remember the email discussions when ECCC was started in 1994. There was a very conscious decision that ECCC papers are “technical reports” and hence are are not exclusive neither of CS conference nor of journal publication. It seems to me that the initiators of the new format consciously decided not make such a decision but rather left the issue open. I would love to see an open non-exclusive interpretation emerge, maintaining the possibility of further conference and journal publication. I would also hope that academic promotion committees when they evaluate such papers will be smart enough to weigh the reputation of the forum rather than the physical format or copyright status, maintaining the prestige and visibility of top conference publications.
Turing's Invisible Hand 
I do not really see how it blurs the meaning of a publication.
If you have a paper in the Proc. XYZ, you have a paper in Proc. XYZ, and that’s it. You will not re-submit the same thing to another conference. You might publish an extended version in a journal.
It does not depend on the physical format or copyright transfers. We already have conferences like STACS with open access web proceedings (which is great) and it does not cause any confusion. I hope other major conferences do the same thing as soon as possible, and we can also avoid problems like this: http://mybiasedcoin.blogspot.com/2009/04/acm-does-not-support-open-access.html
Can a “workshop” (with papers that are not supposed to be new) use EPTCS?
Can a conference that uses EPTCS decide to allow papers that are concurrently published elsewhere? (e.g. there was a time where structures also allowed papers that appeared in stoc, but didn’t publish them in the proceedings.)
Can a “research topic” use EPTCS without having a physical meeting?