Some interesting things are going on in Canadian publishing these days.

As a publishing market, Canada consists primarily of smaller publishers who are highly collaborative, with budgets that are capped locally in a number of ways by central government grant awards and budgets. As with other markets, science and medical publishers have more commercial options than humanities and social sciences publishers -- that is, the publishers who thrive are able to generate business outside of Canada. Consequently, many of them have commercial arrangements with larger US, UK, and EU publishers which help keep the lights on and their businesses moving ahead.

Given its relatively small size, the Canadian publishing market has been the source of a disproportionate amount of news the past year or two, from legal battles over the scope of "fair dealing" (i.e., fair use) for copyrighted materials to the challenge OMICS has posed by locating a major office in Canada.

Now we have the announcement that PubMed Central (PMC) Canada will be closing in February 2018. The last manuscripts are being accepted today (Friday, January 5, 2018). Two reasons are given for the decision to shutter the service:

Approximately only 4% of author manuscripts arising from Canadian Institute of Health Research (CIHR)-funded research have been deposited in PMC Canada by researchers since the system was created
To continue to operate, PMC Canada would require a number of technical upgrades to meet Government of Canada web and security standards. However, the time and resources to upgrade the system are prohibitive.

About 2,900 manuscripts have been deposited on PMC Canada since 2009, when the service launched. These manuscripts will continue to be available through PMC and PMC Europe. This isn't much of a change, it turns out -- the way PMC Canada worked, manuscripts were deposited with PMC (US) and then mirrored to PMC Canada. As infrastructure, PMC Canada was always a bit redundant.

Lack of compliance from researchers has been a long-standing problem for repositories. For PMC, publishers have provided the solution, depositing final manuscripts or finished papers on behalf of their authors. The importance of contributions from publishers to the success of PMC came to light recently in a few venues. The first was when Patti Brennan, the NLM Director, tweeted celebrating the one-billionth article retrieval from PMC. As Brennan notes in her blog post celebrating the landmark:

Current articles follow one of two paths to get into PMC: they are deposited either by the journal publishers or by the authors themselves.
The first path delivers the lion’s share of articles to PMC. Over 2,400 journals have signed agreements to deposit directly to PMC the final published versions of some or all of their articles.

SPARC's Heather Joseph retweeted and amplified Brennan's tweet, ignoring the contribution publishers have made to PMC:

Joseph's "thank you's" go exclusively to the NLM and NIH. She does not add @STMAssoc or @AmericanPublish or @ScholarlyPub or anything similar to indicate the work done by the 2,400 publishers whose staffs reliably deposit, check, flag, and fix PMC deposits on a daily basis. This blindness to the amount of subsidization publishers provide to ensure PMC has papers and functions well is critical to the illusion that PMC is running on government funding, when in reality the government funding supporting PMC is mainly about handling author manuscripts and, more importantly, is just the tip of the iceberg of overall cost PMC demands of the ecosystem.

This blindness to the amount of subsidization publishers provide to ensure PMC has papers and functions well is critical to the illusion that PMC is running on government funding

I saw evidence of the amount of work publishers put into PMC when I had my Freedom of Information Act (FOIA) requests going with the NLM and NIH. One snowy evening, I received two large boxes holding thousand of pages of production manifests between PMC and its publishers. These manifests revealed in excruciating detail the amount of community work it takes for PMC to exist, and that this work is largely done gratis by publishers. These pages represented thousands of email exchanges with publishers large and small about production issues, technical issues, bibliographic issues, and so forth, all drawing on publisher staff time, systems, and processes, at no charge to the US government or PMC. These publishers came from all over the world, with the likes of Elsevier and Taylor & Francis standing out as non-US publishers actively supporting this "national" treasure.

Marie McVeigh, now at Clarivate, responded to Joseph's tweet along these same lines:

PMC is largely subsidized and enabled by publishers from around the world. It only works because of these subsidies. PMC Canada failed to some degree because relying on authors to deposit manuscripts was unworkable, leading to a dismal 4% rate of compliance. It seems that only when professional publishers are doing the heavy lifting does the system work. This is something to keep in mind the next time you use PMC and think it's a great free reference tool funded by the US government. It is there through in-kind contributions from publishers large and small.

Acknowledging the contributions of publishers to PMC and similar industry initiatives recognizes the economic aspects to some services we take for granted. PMC is free because publishers pay to support it, with non-monetary but very real work and expenses. Larger publishers could likely eliminate some expenses without the workload involved in supporting PMC. Smaller publishers might find more time to do interesting new things without the constant burden of depositing manuscripts for authors. It's naïve to think that there is no cost to this support. Just because the costs are hidden doesn't mean they don't exist.

Just because the costs are hidden doesn't mean they don't exist.

The second issue cited by PMC Canada for its closure is the pending expense of upgrading its technical infrastructure, just 8 years after launch. This speaks to the hidden costs of technology, something I've written about extensively before. Not only do systems require constant care and feeding, but upgrades are inevitable. The upgrades to PMC Canada may have been made more urgent by events emanating from the 2014 hacking of Canada's National Research Council (NRC). A presentation evaluating the damage of these attacks set the costs in "the hundreds of millions of dollars," and also showed that hacking of NRC continued to occur from that point forward. In 2015, PMC Canada itself was taken down for a period in the wake of the initial cyber-intrusion.

One can only imagine that with Sci-Hub and its hacktivists probing for weaknesses in government computer systems, any exposed flank would draw scrutiny. (Note that Sci-Hub itself has definitely increased the costs of technology provision in our space, and hackers create secondary costs for security and intrusion remediation.) The upgrades may or may not have been related to increasing system security, but given the history here, it wouldn't be surprising. The fact that PMC Canada was a redundant system with a very low participation rate likely made the decision easier.

The demise of PMC Canada is a cautionary tale about introducing redundant services that provide little additional value and create unforeseen risks.