Discovery tools have been on the market for quite some time now. And while I would dare to say that users seem to at least like these tools better than the traditional catalogs, it seems that these tools go down much less well with librarians. It has been my personal experience in both working on a homegrown solution and in implementing a commercial product that librarians seem to have multiple reservations towards these tools. This is only mildly put, as I found when discussing my experience with discovery-people in Germany as well as internationally.
Being both a librarian and a discovery-enthusiast myself, this has me wondering. I do share many of the reservations that my colleagues have. Whatever we see out there is often far from perfect yet, for instance as far as the actual “discoverability” of large amounts of heterogeneous metadata or the level of integration of discovery systems with ILS, link resolvers and other products are concerned. But for the most part, I think that these are rather interesting challenges on our road to the future. Sure, there are really early adopters of discovery tools like the University Library of Utrecht who has decided to concentrate on delivery rather than discovery. I tend to think of this as an avant-garde decision which could not have been made without the experience of introducing a discovery system in the first place. Putting your own metadata, licensing and availability information in the context of a discovery system (i.e. outside the system this data was born into) and actually making it work there can be a painful experience, but a necessary learning process for all departments in the library.
Is that pain a convincing reason for librarians to dislike discovery tools? This is one of eight hypotheses I have come up with so far:
Librarians don’t like discovery tools because…
- They are too much extra work. As I said before, the level of integration with existing systems is not very good yet (especially when you live in Germany with a whole different landscape) and checking licensing information in the catalog, the EZB (German ERM for serials – sort of), the link resolver AND the discovery system is just too much.
- They weren’t our idea in the first place. Looking at the people who are usually the drivers of decision for a discovery tool and the implementation process, I rather see IT- and management folk than cataloging or reference librarians.
- Strange things happen to our metadata. Our metadata for instance is being mapped from a proprietary format to MARC21. Information does in fact get lost there. Work done by librarians that is already mostly invisible work gets even more invisible. The promise that discovery tools make better use of librarian-created metadata by allowing faceted browsing was not yet fulfilled.
- Talking about the strange things happening to the data is hard. The guys (yes, they are mainly guys) don’t speak PICA, MARC etc. But still they ooze what might be perceived as contempt for the way librarians have designed and are employing bibliographic data. For the record: I get the disgust for this kind of data, because I know where these guys are coming from and I do appreciate the perspective they have brought (and are bringing) into library land. But I am talking about librarians and their perceptions here!
- They mess with the concept of the catalog. A catalog used to be the inventory thing for one library – which makes the bridge between finding and getting items an easy one to cross. Being able to extend searches to other catalogs and bibliographic databases may be what users want, but it surely is a challenge for both discovery and delivery.
- They are hard to use in reference interviews. Librarians know the catalog inside out – small wonder, since it was them who built the catalog. Doing a search in the catalog means getting predictable results, whereas the search in a discovery tool is a whole different matter. Many librarians I have talked to think of relevancy ranking per default as dangerous or even unethical. Not knowing how exactly the ranking algorithm works makes matters even worse.
- They make users lazy and dumb. Sometime ago, I asked people about what they thought where the perceptions of librarians of their users. One answer I got: “they (librarians) think that users should eat their greens”. To put it less succinctly: Discovery causes the mental models that users and librarians have of search processes to clash.
- They cost us our jobs. While I have heard people stating the above-mentioned points in their speeches against discovery, I have never actually heard this particular argument. So this is probably the most far-fetched one: In fact, this thought occurred to me myself once and I am wondering if this resonates with anyone. Like I said, I think that discovery as we know it is far from perfect yet. But if it was? What about the reference desk and the classes we teach?
Neither do I know if these hypotheses are valid, nor how to best examine them further. I pitched the idea of doing that to some people in library school, but they don’t seem to bite yet. I am not very knowledgeable when it comes to research methods on perceptions. I would be also very concerned about any interview/questionnaire/whatever sounding in any way condescending. Because, as ever, I think that even with our weird and strangely-structured metadata, our profession and its virtues can make a real contribution to discovery (and delivery, come to think of it). But I am also under the impression that there is too much unspoken discovery-related agony out there for this to work out on a larger scale – which is why I would like to see this examined and brought to light. But maybe somebody knows of any research that has already been done on this?
P.S. (Feb 5): In the excitement about writing in English, I forgot to mention to colleagues who have given me inspiration for this article: I am grateful for discussions with Anja Knoll about how librarians perceive discovery tools, and for the introduction to the concept of invisible work (a German article mentions librarians specifically!) which I got talking to sociologist Carola Schirmer.
So, e-Science and libraries. Maybe I was having a bad day, but I left the course on on “The Library in the Scholar’s Workflow and Research Data” at the Ticer Summer School with a distinct sense of despair: Not only did it stay unclear to me what the role of libraries and librarians might be. Also I could not help but wonder: Isn’t the search of libraries for new services to deliver in order to prevent their much-feared fall into into insignificance just pathetic? Although I am usually quite enthusiastic when it comes to unusual or time-consuming measures to make a difference, both for libraries and the profession. What happened to me?
At the course, I have seen terrific examples and overviews of virtual research environments (VREs) and learned of projects that try to internationally aggregate research data. But what can we as librarians contribute to that? Of course: We provide metadata on printed material and as much full text as our funding stretches to. We have some expertise in the organisation of knowledge (though mainly in form of printed material) and care about archiving. But does that mean we should take the lead on the building of the actual tools that researchers require in order to collaborate and share? Carole Goble from Manchester gave a demo of a platform in systems biology that allows researchers not only to share data, but also to rate research and thus help to build enough trust to foster even more sharing. But Goble, despite all her brilliant ideas for features of e-research frameworks, seemed hard-pushed when she was asked to come up with ideas for the role that libraries might play here. Andrew Treloar from Australia’s National Data service took the same line in his presentation “Data, librarians, and services”: While very respectful of the knowledge of librarians regarding metadata creation, he felt that the librarians’ ideas for metadata schemes and interface design were too complicated and “purist”.
I am not saying that the library profession should not care. Big libraries like the British Library or, in Germany, the TIB/UB Hannover have initiated Datacite, an international consortium for the collection and archiving of research data, for which they have just been rewarded with an award for “Rethinking Resource Sharing Innovation” (Congrats! Seriously!). But the average subject specialists will probably have neither enough time nor know-how to start leveraging e-research, even if they did have excellent contact with the communities they serve. And even if a local group of researchers got the ball rolling and approached their library for support in the VRE-building, would the library be able to give them want they need?
- The content libraries offer is oftentimes not ready for re-use. Catalogs, repositories and other databases lack APIs which would be needed for an automated integration of metadata and fulltext into the VREs. And we are very reluctant to give up on our established metadata schemes and models (not to speak of licensing models). So before we can start playing the field of e-science, maybe we should ultimately decide to open up the silos that are our information services and prepare to let go of whatever “purist” views on metadata that we might indeed have.
- Librarians are trained to gather and collect information. But I suspect many librarians would feel uncomfortable to be cast into the role of content curators or even arbiters, which is actually one of the few roles that researchers seem to have in mind for us when it comes to VREs. We are all about neutrality. Our subject headings describe content in the most non-judgemental fashion. Are we prepared to make decisions about which datasets to archive?
I am particularly concerned about the lack of openness of our information systems, both in a technical and conceptual sense. If we don’t fix this, I don’t think that engaging in e-science is going to make much of a difference, let alone be a safeline for libraries who would otherwise just become the Starbucks with books. On the other hand, it would certainly be inspiring to see a few best practices for e-science projects with significant involvement of librarians and library content. Seeing as some studies seem to indicate that e-science is going to evolve slower than expected, there might be time to prepare – and in spite of all the despair, this is pretty much what the course was all about.