Archive for February 1, 2019

The Cost of the Open Journal of Astrophysics

Posted in Open Access with tags , , , on February 1, 2019 by telescoper

Our recent publication of a paper in the Open Journal of Astrophysics caused a flurry of interest in social media and a number of people have independently asked me for information about the cost of this kind of publication.

I see no reason not to be fully `open’ about the running costs of the Open Journal, but it’s not quite as simple as a cost per paper.

The Scholastica platform we use (which is very nice, simple and easy to use) costs $99 per month. That includes professional website hosting with a custom domain, a built-in website editor (so the site itself can be easily customized), integrated PDF viewer, indexing through e.g. Google scholar, fully searchable metadata, and readership analytics. That amounts to $1188 per annum, regardless of how many submissions we receive or how many articles get published.

On top of that we pay for the Peer Review service, which amounts to $10 for each submission (subject to an annual minimum of $250). We pay that whether or not a submission is published. So far we have rejected significantly more than we have accepted. This system provides automated emails, deadline reminders, an interface for searching sorting and assigning submissions to editors, file versioning & blindness control, a reviewer database, metrics to track performance, etc.

The final charge is only for papers that are accepted: we pay a fee to Crossref to register the Digital Object Identifier (DOI). That costs a princely $1.

There are no* other charges, as editors and referees give their services for free. Since all papers are typset by authors we do not use the Scholastica typesetting service (which is $5 per 500 words). If you’re thinking of setting up a non-overlay journal you might want to pay for that.

The actual cost per paper therefore depends on how many papers we publish. If we had 25 papers submitted in a year and published 10 the net cost per published paper would be ($250+$1188+$10)/10= $144.80, but that reduces as the number of published papers increases. For 50 submissions with 20 published it would be ($500+$1188+$20)/20=$85.40, and so on.

Some publishers argue that Open Access publication justifies an Article Processing Charge of several thousands of dollars. I think I’ve demonstrated that it doesn’t. Any charge over a hundred dollars or so is pure profiteering, bearing in mind the huge economies of scale inherent in large organizations.

In reality we have a combination of sources of funding that will be able to pay the annual fee for the foreseeable future. Ignoring this element, the marginal cost per published paper is just $11…

I hope this clarifies the situation.

*As has been pointed out in the comments, there is of course the cost of running the arXiv. The current funding model for that involves a membership program according to which institutes pay a fee depending on how heavily they use the arXiv. The top fee is $4400 per annum, for an entire institution. Some OA journals charge that much as an APC for a single paper!


25 Years of Python!

Posted in History with tags , , on February 1, 2019 by telescoper

Not a lot of people know* that it is 25 years to the day since Guido van Rossum announced the release of Python 1.0.0:

The latest version of Python is 3.7.2.

It’s not quite correct to say that Python is 25 years old today, though. There were versions available before the official Version 1. For a full history see here.

*H/T to Tom Crick, whose tweet alerted me to this.