Random musing on the ‘headless library’ and other search trends …

This year I’ve spent a lot of time considering the perceived value of library search services against their actual use. Running and improving library search services has been the core of my job for the past few years, but recently I’ve been involved in discussions on search with colleagues from a non-library background. I’ve spent time explaining (often seemingly justifying) why we’ve taken the routes we have.

I can understand their point of view. For those not familiar with the way libraries work, search is a ubiquitous commodity and normally done via Google, Wikipedia and Facebook. We’ve developed our own parallel sets of data standards, tools and approaches to the rest of the world and are involved in something of a course correction.

Why do we need to be different, and if we don’t, what does it mean for us?

I’m still running this one through in my head, but here are a few random points:

  • We evolved our search services from OPACS and A&I indexes. These were gateway services when information was scarce, but now-days, information is everywhere and librarians are committed to a shoulder-to-shoulder mode of engagement
  • Three to four years ago, Tony Hirst was telling me (and others) that library catalogues and discovery services are rarely used to actually find material of scholarly value. He was not alone in this arguing that discovery actually happens elsewhere, be it on a search engine or via a reading list of citation. So I do wonder, are our discovery services actually access services? Handy for dishing up proxied links to online resources and shelf locations and class-marks to books, but not for much more? Is the information need they meet need not one of ‘what’ but of ‘where’?
  • Does accepting that discovery happens elsewhere mean totally giving up on our own search services? At least one library has postulated this option, dropping its own search engine and educating and guiding users into getting the best of the tools already out there. It makes sense from one point of view, these are the tools users will use every day. This is especially true for federated search. No-one in the world outside of a University is going to use anything remotely like Metalib anymore. Thankfully. You could call this approach the ‘Headless Library’, one focused more on assistance, disentanglement and fulfillment.
  • So if we all did this, what do we loose? What about the role on a public inventory? Search ‘one stop shops’ are a misnomer,  but a useful first stop shop or a store front for the library’s’ collections via the web are still valuable things to have, if only to show the value of a libraries investment in its collections.
  • Apart from anything else, search engines cannot yet deliver class-marks or proxied URLS. Known item searching to give appropriate access is still a valid use case, but does it require a whole complex faceted discovery service? Especially one you need a training course in to fully understand …
  • There are other things we can loose with this approach. Although Google and Facebook are adopting complex underlying structure, academic use cases need certain traits not found in these tools. This includes fidelity and granularity, searching over specific metadata fields and terms. I would argue we also need some degree of transparency, highlighting how a search result was found. For some use cases, such as literature reviews search services need some chance of repeatability. Against these criteria, even some library centric tools may come up short. Many of these terms are also core values for Information Professionals themselves, worth hanging on to.
  • Even thought they wow us with new features, its always worth remembering the commercial imperatives behind search engines and social networks. Even if we got all of our data into there, booksellers and ebook sites will always be valued over library records. Our niche cases like special collections might shine, but search engines are not great for known item searching for library held textbooks, (although users may be lead to cheaper second hand copies on ebay that are easier to get to than a library copy …)

So why are our users going elsewhere to discover? Quite possibly because search engines and subject-specific resources adequately meet users needs already. This may be a fact of life we just have to adapt to. It may be that discovery for them actually involves a social interaction, a recommendation via a peer, citation or reading list, not even a search.

There may be other reasons, Rick Andersons’ well argued concept of surplus value in libraries springs to mind. I’d argue we are massively guilty of adding too much polish or helpful links to our interfaces. We should always remember that every feature or link we add incurs a debt. This can make our search engines much harder to use, even if we are just trying to be helpful. Its great to see a growing trend in UX Design in libraries specifically aimed at turning this ‘over polish’ around. This movement is more than a fad, its also getting its own journal.

Lastly, if you do want to see a beauty of a catalogue (and I still do), then check out the National Library of Finlands’ customised VuFind implementation. The drag and drop date selector is great, as is the tight integration of archives into results. A great example of how strong design and a clean interface can work with structured data. This is a genuinely brilliant national level resource and great way to showcase the value of a library service.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s