Funding agencies should really stop kneeling in front of publishers and refuse to pay for this, or at least fix themselves a maximum price per publication. Currently publishers ask for $1000 to $5000 in some cases, which is completely insane. I'm not denying the cost of publication. But there is that actual cost, and then there is the prices demanded by publishers, which looks more like violent theft than it looks like the actual cost.
I believe copyediting and layout is paid, however.
The major publishers (who need to get involved to have the proceedings indexed and "counted") get everything ready to publish online (well, they do slap a cover on the thing and get an ISBN) after the review process and technical process is done, and still try charge for access to the articles.
I've published in Nature Physics, Physical Review Letters, PRA ... and we did 99% of the copyediting and layout (with LaTeX and unpaid, of course). The editor would correct a few typos, modify the layout slightly (1h of work tops) and that was that.
Some open access journals charge authors. But these open access journals are not the publishers we hear about (such as Elsevier). In some ways, it's accurate to say that the old model with publishers is "reader pays", and the open access model is "author pays".
The good thing about open access journals is that they are typically not a for-profit endeavor. This means that what they will charge to the authors is, hopefully, reasonable.
I'm insisting here on the fact that publishers attempt to melt the notion of "open access" and "gold open access" with author processing charges, but this is manipulative propaganda (and sadly it works, lobbies at EU and in the US succeed in convincing politics and funding agencies that they have to pay to publish in "open access").
Sources:
http://www.plos.org/wp-content/uploads/2014/09/about-journal... http://www.biomedcentral.com/about/apcfaq/howmuch
Now I think you are thinking about this the wrong way. Nobody is denying that publication has a cost. But take all the money that universities around the world spend in subscriptions, add all the money that is paid in APCs. With that kind of money you could laaaaargely support many publishing platforms and professional librarians and technicians to manage, improve and maintains articles metadata and the platforms.
The remaining amount of money (which would be most of it in my opinion) could be used to actually fund research and higher education.
But let's be even wilder! Consider that, rather than being owned by a publisher who's main job is making money, each journal is owned by its editorial board (i.e., the scientific committee composed by the researchers who actually do the peer reviewing work for this journal as part of their researcher work (they're not paid for this task in addition to their salaries)). Now with relatively cheap¹ technical platforms (such as arXiv, HAL, revues.org, easychair, sciencesconf, SciELO, …), we would have everything we have now² for almost free. And publishers could act as contractor (and would be in competition against each other for quality and price) when we want professional editing and paper prints of some journal volumes. If in addition to that the papers use free content copyleft licenses such as CC-BY-SA, then when any university choose to spend a little money to get a professional edition of some papers everyone benefits.
¹ These platforms are not that cheap in fact, but they cost nothing compared to the current cumulated spending in subscriptions etc.
² Many publications todays are actually not professionally edited, publishers actually take the camera-ready version of the authors and just put the PDF online. So their job mainly consist in hosting PDFs, yet they get a freaking awful lot of money for that.
I point to PLoS as essentially the model you're asking for. It's a non-profit publisher started by open-access advocates working as researchers. There are no investors making a profit and they're mission is to spread open-access as a model, and they end up charging $2200+ per article for most of their journals. If that's the number they came up with, then I'm led to believe that that's what it costs to publish a journal, at least in the biosciences. As I said, there are small journals that operate for free, but if you look at large-scale open-access operations, started by idealists, I don't think you're going to find any that sustainably gets by with article charges less than around $1500.
For comparison, when I looked at Elsevier's finances, it seemed that they're charging roughly $5000 per article and half of that goes to their operations and half is profit. So their costs are roughly the same as PLoS, but they charge twice as much in order to make a substantial profit.
As for why I felt the need to correct you, I think it's because I think in general people ignore the real monetary costs associated with open-access publishing. I believe that author-pays open-access is the way of the future, but getting there will mean that researchers will have to find ways to pay these fees when they haven't had to pay before. In principle, this just involves redirecting money from subscriptions to pay author charges and there'd even be money left over, but actually performing that switch across thousands of institutions is not simple. Too often I run into the attitude that for-profit publishing is charging money for a trivial job, and so people wonder why it hasn't been replaced. However, the truth is that they're performing an important and real task with real costs, while also over-charging for it, and that's not going to be replaced overnight.
Of course neither me nor my lab could afford that so I just went with the regular way, signed their fucking copyright release form while cursing a lot against them, and in the end didn't really care because in my field (cryptology), most of the papers are on the IACR ePrint Archive even before they are published (including mine) and they are updated there when they change.
So, my personal experience and my research on the subject (I have done a lot of work on open access, see http://pablo.rauzy.name/openaccess/introduction.html which is only available in French for now) say that no, the publishers work is most of the time not that important: for instance in my case if the conference or journal just published a list of links to the accepted versions of the paper on the IACR ePrint archive, the cost of publishing would decrease drastically and the quality of the publication would not change at all.
How many articles do they publish per month and what are their costs?
We expect peer reviewers to work for free. We expect the authors of the papers to work for free. But no, we don't expect the publishers to work for free.
As p4bl0 wrote, "But there is that actual cost, and then there is the prices demanded by publishers, which looks more like violent theft than it looks like the actual cost." That means p4bl0 think the prices should be reduced, not that the prices should be zero, as you somehow concluded.
"Go publish it on your blog."
See, I've actually done that. I can tell you what happens. A large subset of people don't believe something has been published unless it's in the scientific literature. People end up writing papers which say things like "the first comparison between X and Y". I write to them pointing out several previous blog publications which report the analysis, to be told that they knew about one of them, but didn't consider it sufficiently academic to count as a real publication.
Publishers make their profits because a lot of people, including those who hire or fund academics, believe that a publisher's imprimatur is essential to being a publication.
This is not completely invalid. Google's standard search engine, good as it is for search, is lousy for searching the academic literature. Google Scholar is better, but it only manages it because it only indexes the academic journals.
So if you publish on your blog, it's not likely that people will know about your work.
If you want to advance career in science that has a strong academic orientation, then it's very hard to do so by only publishing blog posts.
Therefore, if you don't like it, change the system. Which is what this debate has been about for the last 15+ years.
One paper I'm thinking of compared implementations of two different algorithms, and showed that for chemical data VF2 was several times faster than the Ullman algorithm for subgraph isomorphism. A couple of years earlier, two of the free software projects in the field did the same analysis, with the same conclusion. Both published their results on their respective blogs/wiki, and changed the internals to use VF2.
These don't resemble a publication in a way that Google Scholar can discern.
It's a bit annoying to me in that the scientific literature is supposed to be "self-correcting", in the sense that I could publish a followup paper highlighting some of the pre-history. But the journal I'm thinking of is an OA journal, with no letters to the editor or similar section. The only way to update the literature is to pay ~$1,000 for a full-sized paper, or convince some other journal (... or arxiv? Hmm...) to publish a correction piece.
Quite annoying.
Google scholar seems to index arxiv.org (a free for readers and authors, no review paper archive). Would you consider publishing those things you've published as blog posts there too?
For example, under Quantitative Biology you'll notice a lack of pharmaceutical chemistry, toxicology, or pesticides, which are closely related areas to the types of work I do.
Searching out of curiosity, I found http://arxiv.org/abs/1303.1724 as an example of something in my field. It's under the catch-all "Computer Science > Computational Engineering, Finance, and Science", so that's a possibility.
There is some toxicology under metabolomics, but that focused on expression data, not structure-based prediction. There's another tox paper filed under "Mathematics > Statistics Theory", again based on gene expressions.
There's no history of people using arxiv.org in my field. If people don't like to cite blog publications, I don't know if they would like to cite preprints. Something to find out the hard, slow way, I guess.