The “blog and secretaries” model of scholarly publishing

Steve Harnad has done much for open access. So I was surprised by a new article, in which he wrote:

In the old pre-digital days of S&S (“scientists and scholars” - ZF) publishing, the true costs of providing print-on-paper to would-be users required the services of another profession for the production and delivery. But (let’s cut to the quick) those days are over, forever. Online publication is not altogether cost-free, but the costs are so ridiculously low that all an S&S author needs pay for is a blog service-provider, rather like a phone or email service provider. ...

All other goods and services are now obsolete – or almost obsolete: scaled down to only the tiny, trivial cost per article of managing the peer review, by paying secretaries to run the software for soliciting and monitoring it.

This is such misleading characterization of what is required to run an online publication that I kind of can’t believe he wrote it. I think he knows better – I hope e knows better – but is oversimplifying to try to make a point.

Before I go on, I want to be clear. I am not about to argue here is not that journal prices are fair, or they could not be lower. When I commented on this article on Twitter, Björn Brembs came in and provided what the article lacked: some real values associated with costs. 

What I want to comment on is Harnad’s characterization of the infrastructure of an online publishing platform. A blog service and a secretary is pretty much all you need, according to him.

In the last decade, there have been no shortage of people providing online scholarly platforms. Many of those are trying hard to move away from legacy academic publishing and the costs associated with those traditional journals.

Very few have been able to make the “blog and secretary” model work. There are possible exceptions, judging from this description of the Journal of Machine Learning Research. Keep in mind, though, that this description is now over ten years old. I am sure there are other niches for the “blog and secretaries” model, but I don’t think this is the way forward for all scholarly communication.

Preprint servers provide open access for readers without fee to authors. They are no legacy publishers, and would seem to be prime candidates for Harnad’s proposed ultra-lightweight publishing model.

This article describes how several preprint servers were struggling to get enough money to run. Since the article appeared in 2020, a of them – INA-Rxiv, ArabiXiv – are no longer taking new submissions.

Brembs mentioned Humanities Common as a model. It has expenses of about US$750,000 and is run by three major committees. This still seems a far cry from the picture Harnad paints of using off the shelf software and a skeleton staff.

Likewise, larger preprint servers have higher costs and a bureaucracy to sustain them. The big one, arXiv has a budget of US$4.8 million.

I would love to survey of researchers on their estimates of the costs of a preprint server and how many people it takes to keep them running. Prediction: I bet a lot of them would underestimate the cost and time it takes to do so, perhaps influenced by descriptions like Harnad’s.

External links

Publishers can’t be blamed for clinging to the golden goose

Popular preprint servers face closure because of money troubles