[No 2 of 3. No 1 here.]

I was pleased to participate in LITA's Top Tech Trends panel at ALA this year (see the video and live coverage).

We were each asked to talk about three trends: current, a bit further out, and a bit further out again. In thinking about the exercise, it seemed to me that it would be interesting to talk about how services are being reconfigured in a network environment, and not just focus on technology as such. This is the second of three blog entries, one devoted to each of my trends. We had three minutes in which to discuss each trend.

I really only decided to talk about my second trend the day before the event. I was influenced by discussions with several people as I wandered around the exhibits hall. Opinions varied as to how important this trend is, but I chose to talk about 'discovery layers' because it seems to me that if these become successfully and commonly deployed they have quite far reaching implications.

What do I mean by discovery layer? A discovery layer provides a single point of access to the full library collection across bought, licensed and digital materials. Typically, a single search box is offered alongside a range of other navigation features. Products which support this approach include Worldcat Local, Summon, Primo Central, and the Ebsco Discovery Service, as well as a range of institutional, national or other initiatives.

Working with hindsight ;-), these are the points I meant to make ....

  1. The full library collection. If they develop as anticipated (a real question), the discovery layer will become the view of the library collection for library patrons. In fact, for many users it may actually become the library. This has several consequences:
    • What is not represented in the discovery layer will be much less visible.
    • There will be pressure to incorporate more services into the discovery layer - better fulfilment for example through resource sharing, Google book search, purchase or other options.
    • The integrated discovery experience will more clearly expose lack of integration with services behind, and will drive greater integration. One can see, for example, potentially more interest in the direct-to-content approach of something like PubGet.
    • And as somebody suggested to me afterwards, there will need to be strategies for managing those who resist the loss of a specific database interface.
  2. A driver for other operations. If the discovery layer becomes the central focus for access to collections, then one can imagine discovery patterns begin to affect supporting operations like selection and acquisition. The patron-driven acquisition model is being explored in the ebook market - will it be extended to other licensed materials?
  3. Data wells and the provider landscape. A discovery layer depends on an aggregation of data - a 'data well' - which involves considerable coordination costs. These include the processing involved in normalizing the data and the business interactions involved in assembling the data. The level of normalization may vary - how much work, for example, do you do in clustering author names across A&I databases, catalogs, and so on. It does not make sense to do this work too many times, so one might expect a small number of providers to emerge who syndicate 'data wells' to others as well as use them in their own services. It will also be interesting to see how strong the tendency is to use other products from your discovery layer provider - a knowledge base in which to record licensed holdings, a resolver, and so on.
  4. Indirect discovery. It is important to remember that a discovery layer 'destination' is a part only of the library user's discovery experience. Increasingly the library needs to think about how its services are visible to users who discover their information resources in Google, in the course management system, and so on. I discussed some issues in a recent post.

Comments: 4

Jul 05, 2010
Till Kinstler

I agree, that discovery layers expose the lack of integration with existing library services. That's one of the main topics in library technology for the coming time, I think. When you look at the ways, current discovery solutions integrate with the currently existing library services infrastructure, that's really ugly from a technical point of view (and often from the user's point of view, too).
There are some inititives addressing that, most notably the ILS-DI working group (http://groups.google.com/group/ils-di).

Though, from a technical point of view, I doubt, discovery layers drive more technical(!) integration (I don't think you wanted to say that, because the article addresses discovery layers from a non-technical point of view, so I just want to add my technical point of view).
On the technical side, we need disintegration of those integrated library systems to connect discovery layers and other software to the existing systems and services infrastructure. We need open interfaces (APIs).
We need choice. We need that to provide users with an integrated view of the library services and thus better service (see websites like those of the ETH Z├╝rich library or University of Michigan library for integration of discovery layers and the complete library web site in one user front end). We need that to finally bring libraries on the Web...

Jul 05, 2010
Frank Winter

Lorcan, could you expand a bit - either here or perhaps later - on what you are alluding to when you say in your first point about the full library collection,"if they develop as anticipated (a real question"? There are identifiable risk factors (and many different kinds of risk, of course) with any projection for developments of any of the collection areas of the OCLC grid, for example, as well as unforeseen risks as always, but has there been any discussion of alternate futures?

Jul 09, 2010
Stephen Francoeur

As excited as I am about discover layers, I do worry deeply worried about those databases that don't work in this environment (due to technical limitation or licensing restrictions). Students and faculty at the college where I work depend on resources that are data driven that are unlikely to find a home in discovery layers; I'm thinking about things like Thomson ONE, Reuters terminals, Bloomberg terminals, Wharton Research Data Services, Datastream, etc. Where you say, "What is not represented in the discovery layer will be much less visible," understates the potential downside we may see. These products are among the most expensive resources we license; their absence from the library's "search engine" (as our students are bound to call our discovery layer) will prove painful to us in many ways. Is there much hope for including these data-driven resources in discovery layers?

Jul 26, 2010
Lorcan Dempsey

Frank - what I was getting at was that discovery layers may not take off in the way that some are suggesting because of various of the issues around them. Can they be comprehensive enough of licensed materials? Can they synchronise effectively enough with local repository or catalog systems? Can they provide enough customization or API support? And so on ...

Leave a Comment