Martin Weller has got a bit irked by someone likening open access journals to stealing bread, and makes the obvious-to-us points:
Open access is not like stealing bread from Tesco for two VERY SIMPLE reasons:
i) If I take a loaf bread from Tesco then it is no longer there. If I read an online article it is still there for everyone else. We knew this back in 1998 when we talked about non-rivalrous goods. Can we stop using this misguided analogy now? Pleeeease?
ii) You could make the argument that each download of pirated music equals a sale lost (although that is a flawed argument I think). So maybe the argument has some validity for products you want to sell (big maybe for me). But for journal articles it really has no validity – the whole point of research publications is that you want as many people to read it as possible.
I agree, although there’s some interesting stuff buried in point (ii): for academics, researchers and the General Public Good, the whole point of research publications is for as many people to read it as possible – or more precisely, for everyone to whom it is relevant to read it. But for a for-profit company, the whole point of any activity is to make money for the shareholders. The happy part of the history of capitalism is that generally, the most sustainable way of making money for shareholders is to do things that people value. Without getting too far in to the whole critique-of-capitalism argument, if restricting access makes more money for a commercial publisher, then the company has a legal duty to do that. The interests of academics, researchers, the General Public Good and those of commercial publishers were very happily aligned – or at least, very usefully overlapping – for perhaps a couple of hundred years. I think that this is ceasing to be the case, now that the marginal cost of publication has fallen to nearly zero.
That last bit warrants a little unpacking: the marginal cost of publication has fallen to nearly zero. The marginal cost of publication is the cost to produce one additional copy. That’s not the same as the average cost, or the cost of producing the first copy. And the cost has fallen to nearly zero – not actually zero. There is a difference, and it can be a very important one in some circumstances.
The difference between “the cost of publication is zero” and “the marginal cost of publication is nearly zero” can sometimes be negligible – but sometimes it can be profound. And, of course, the price and who pays it is another matter altogether, or at least can be. This deserves further exposition in the context of scholarly publishing, but this margin of my time is too small to contain it.
Anyway. Martin concludes:
“It’s not as if digital content is new now is it?”
Indeed not. I was just thinking that this morning. I think of myself as working on new technology in teaching, and when I came to the OU (1998) you could use “web-based teaching” fairly interchangeably with “new technology in teaching”. It’s largely the web which has driven the marginal cost of publication down to nearly zero, and the changes were well afoot even then.
But the web isn’t really very new these days. The web got going in April 1993, when CERN said the WWW would be free to use and Mosaic was released. People born then are sitting their AS levels right now, and hoping to go to university next year.
I like to use 1993 as the start date for the web because (a) it’s when it started to catch on, and (b) it’s when I heard about it, which is obviously a key moment in any technology’s lifecycle (!). (As I like to point out before issuing any technological prognostications, my verdict at that point was that it was rubbish and would never catch on. I have since changed my mind.)
If you want to be purist, it’s even more stark. You could date the birth of the web to when Sir Tim had the first complete set of web tools working (browser, editor, server, pages): Christmas 1990. People born then are going to graduation ceremonies right now.
It’s easy to overstate and oversimplify the case for a Net Generation – as my colleague Chris Jones and his team will tell you. But I don’t think it’s going too far to say that it’s hard to call a technology entirely new if it’s been around for a generation.
Doug Clow's Imaginatively-Titled Blog 
Worth reemphasising that it’s the marginal cost not the cost that’s nearly zero as you say this is an important distinction. I worked as a copy editor for a quality mathematical journal based upon being pretty much “as free as possible” (open access but charge for subs and paper copies) and it was well respected. However, it takes time to get a complex mathematical paper into journal style and ready for publication even when the source is good. I was paid by the hour and I didn’t milk it and I can still think of a single paper I spent 13 hours on (and that’s with the LaTeX source) getting it into house style and ready for publication. I was paid for my time of course.
Oh yes – copy editing is a non-trivial cost, especially if you have anything more than just plain text. And even then.
It’s easily the second least fun thing about being Editor of JIME (after ‘chasing people to do stuff’), and my co-editor Patrick McAndrew does almost all of it. And a cursory close examination of our back issues will show you that we don’t aspire to perfection. We’re currently working on a new system (based on OJS) to replace our creaking broken one, and just processing the back issues in to consistent XML is taking something like an hour or two per article. (I think we need to change tack and pseudo-automate the process, but it’s not trivial since the data aren’t consistent.)
An even bigger cost is the writing of the paper, the big-honcho-Editing (initial screen, appointing referees, deciding what to do when the reports are in), and the Reviewing/Refereeing.
And, of course, setting up even a very basic online journal management system (even these days where it’s relatively straightforward) and hosting it and all that costs real money too. Not stellar quantities but definitely not zero.
Given all that sunk cost, the extra cost if J. Random Researcher downloads one of our papers is pretty tiny. But we’d certainly notice the cost – in time dealing with the technical fallout, at a minimum – if the site got Slashdotted.
Good luck with the automation. Hopefully it will save some of the workload. I guess the discipline from the copy editing did help in my own paper writing. Pick a set of style rules and enforce them.