Issue 54, 2022-08-29

Editorial: On FOSS in Libraries

Some thoughts on the state of free and open source software in libraries.

by Andrew Darby

I’ve been thinking recently about open source software in libraries, partially because I was wondering how I could squeeze into this editorial a plug for SubjectsPlus [1], but also because I’ve noticed, in an unscientific way, that libraries seem more closed to open source. When I started in the library biz, over twenty years ago, we talked about open source software a lot. It had a feel-good glow around it: it was righteous, it was collaborative, it was free! It was accessible to the self-credentialed and the tinkerers.

Installing free and open source software (FOSS) was a way for people in libraries to concretely skill up: a cataloger or a reference librarian might muck around with setting up the LAMP stack, encounter a few install errors, and eventually move on to detailed bug reports and then pull requests. Their career might veer off in an interesting new direction. Learners everywhere–intentional or not–need a concrete project to work on, and FOSS applications can fill that role.

FOSS used by libraries consists of both library-specific projects (Access to Memory, Blacklight, CORAL, DSpace, Evergreen, Islandora, Koha, Omeka, PKP Project, Samvera, SubjectsPlus, Suma, Vivo, VuFind, etc.) and more general audience applications that are frequently used as a general purpose CMS (Drupal, WordPress) or hidden in the stack (Apache, Elasticsearch, Fedora, Solr, etc.). There are some library projects that are clearly multi-institutional, like Samvera, but the majority have a de-facto host institution.

So why, when interviewing candidates for a library technology position, do I sometimes ask the question: What are the pros and cons of open source software? Cons? What?

* * *

The answer, partially, is that some of those tinkerers got credentialed and/or became managers. They started to think about budgets, and the cost of staff time, and competing priorities. They got pushback from campus IT on supporting unfamiliar applications. Some decided, well, let’s just write a check. For each of the library FOSS projects I listed above there is at least one paid alternative, and it is generally in the ascendant. [2]

Meanwhile, over the years the coding ecosystem has gotten much more complicated. At the University of Miami we are currently in the final stages of rebuilding SubjectsPlus from the ground up. What was once basically vanilla PHP, SQL, HTML, CSS and JS, is now a pyramid of frameworks: Symfony instead of PHP; Doctrine instead of SQL; Twig instead of HTML; SASS instead of CSS; React instead of JavaScript; and for good measure API-Platform instead of a bespoke API. Some space on a server has become a dockerized web app in Azure.

Not only do the learning curves of all these frameworks make it harder for the curious to explore and tinker with a FOSS project, it also means you can never fully know your own application; there are just too many dependencies downstream. That can seem like a security nightmare, unless you have a measure of trust in the countless people out there who have their own passion projects, who leave their code open for inspection, who plug holes when they encounter them. The visibility of open source code is also one of its strengths, from a security perspective: it’s not a black box.

If one can use search trends as a metric of interest or use, “open source” seems to be trending downwards. The first issue of the Journal was published on December 17th, 2007, and you can see that during its lifespan there has been a pronounced decline in (googling) “open source”:

Graph of search terms 'open source' over time Figure 1. Google Trends for “Open Source,” December 2017 – present [3]

But what about in the Journal itself? You could say that open source is in the Code4Lib Journal’s DNA: we are an open access journal (not quite the same thing) which publishes our articles under an open source license (US CC-BY) and recommends that any included code be given an open source license of the author’s choosing.

When I search the backend of the Journal (which is the open source WordPress CMS), I get 254 hits for “open source,” which means that almost half of our 516 published posts have this phrase in them [4]. In fact, each of our 54 issues has at least one mention of open source in it. Issues 32 and 40 had only a single hit, while issue 12 had an impressive 10. This editorial alone has hundreds of hits!

So let’s bow to the numbers: open source is alive and well in libraries. At least with the people who write and read about it.

Notes

[1] https://github.com/subjectsplus/SubjectsPlus/ I had hoped to be doing a plug for the new version of the software, but with the departure of a key developer, we’re doing calisthenics just short of the finish line. Version 5 coming soon-ish!

[2] My own project is no exception; we have gone down from over 75 libraries worldwide to, well, I don’t want to count. This could be a factor in my impression that libraries are souring on open source. Or it could be that folks are just waiting for Version 5 of SubjectsPlus to switch back.

[3] https://trends.google.com/trends/explore?date=2007-12-17%202022-08-19&q=%22open%20source%22

[4] And if one were to broaden your search to include articles that utilize open source applications (but without using the phrase “open source”) or that share code inline or in an external repository, the number would be considerably higher.

About the Author

Andrew Darby is the Head of Web & Application Development at the University of Miami Libraries, a member of the editorial committee of the Code4Lib Journal and, you guessed it, the founder of the open source library CMS SubjectsPlus.

Leave a Reply

ISSN 1940-5758