INSTG004 lecture for UCL DIS students - Discovery at the University of London

Editor's Notes

• #3: I work for the University of London at Senate House. This is the central administrative body of our federal university of 18 colleges and 10 institutes.Senate House being built in 1937 from University of London archives.
• #4: Senate House Libraries (plural) contains: Senate House Library,a spectacularresearch library of 3 million items concentrated in the arts and humanities and some amazing special collections. The specialist research libraries of the School of Advanced study – institutes of the central University of Londonadding another 1 million items. Our librarysystem includes cataloguesof Heythrop College, University of London, and the library of a nearby art museum (the Wallace Collection) and our own University of London Institute in Paris.We support research and teaching for our federal University of Londonas well as for researchers from about 1000 universities around the world.
• #5: This is our schedule for today. I’ll talk for an hour about the role of the systems librarian and our project at Senate House.Then a break, then we have two practicals where you’ll look at examples of next-gen systems.
• #6: Something about my role and what I do. I’m drawing on experience at various libraries I’ve worked at and not just Senate House.I’ll pick out three key areas.
• #8: Some definitions: OPAC, OPAC 2.0, and discovery. I would normally use the term “next-gen catalogue” to include everything that is not an old fashioned OPAC – I would lump together OPAC 2.0 and discovery layers.In this example I will refer to Innovative Encore as an example of an “OPAC 2.0” type of system.I’ll resist the urge to explain this all by way of a Venn diagram… Where I draw the line between OPAC 2.0 and a discovery engine is discovery should make use of a central index or knowledgebase of data including electronic journals and other subscribed content.They use a big local database for speed, so you get a search experience somewhat like Google search.
• #9: We implemented a new system in 2009/09 but didn’t make it live until 2010.Our old catalogue had not changed very much since the 2000s - the product is a traditional web OPAC from Innovative, called the WebPAC.Our senior management team identified a need to bring the catalogue up-to-date and provide a system that behaves much like web sites our readers are used to using elsewhere.They also specifically picked out a wish to include Web 2.0 features such as user tagging.
• #10: We positioned Encoreas the default,main catalog thatwe direct readers to start from. Briefly wherever we used keyword search previously we are now driving traffic to Encore. This includes sending searches from the WebPAC front page.At the time we launched this we judged from other Encore sites this was reasonably aggressive. Nowadays less so as more libraries are launching a next-gen catalog as their only or their preferred catalog.To drive reader uptake of the new catalogue I wanted to offer Encore as the default option on that places that really matter to us – on the Senate House Library homepage and on the old catalogue homepage. It was important that by following the path of least resistance readers would end up with the new catalogue.This is based on: Experience of other UK libraries who launched alongside an old OPAC and found weak positioning lead to little takeup. Related to that is Marshall Breeding’s argument that this positioning has a marketingfunction.The preferred catalog should have top billing to emphasize its use and ensure we get feedback on it.Breeding, M. (2010). Next-gen library catalogs. London: Facet
• #11: This is the library homepage at http://www.senatehouselibrary.ac.ukWe have a search box positioned prominently on the front page. A fair amount of traffic goes via this. This uses Encore only and for technical reasons can’t offer very many search options.
• #12: This is the old WebPAC catalogue front page. Historically lots of search traffic came in this way – many of our readers understood this is the place to start searching “the catalogue”.What’s happening in this box up here: if you keep our default keyword search (called ‘Quick Search’) you are directed to the new Encore catalogueand the search is pre-faceted by location (or if you select a different index you go to the WebPAC – the old catalogue.http://catalogue.ulrls.lon.ac.uk/search~S1There are sound reason for doing this, mainly drawn from experience testing Encore and feedback from readers following deployment.
• #13: Here’s the resulting search engine results page in Encore for ‘paranormal’, you can see a location facet or scope has been pre-applied to my search.http://encore.ulrls.lon.ac.uk/iii/encore/search/C__Sparanormal__Ff:facetlocations:ull:ull:SENATE%20HOUSE%20LIBRARY::?lang=eng&searchtype=Y&searcharg=paranormal&searchscope=1&SORT=D
• #14: Our Kuali OLE projectmakes things interesting as it does not come with an OPAC as a module. This is a very deliberate strategic choice from the Kuali Foundation partners developing OLE.OLE is a next-generation ‘library services platform’ or ‘unified resource management’ system similar in outlook and scope to systems like Ex Libris Alma and Innovative Sierra.Marshall Breeding on the concept:“To make up for functionality absent in their core integrated library systems, many libraries implemented a cluster of ancillary products, such as link resolvers, electronic resource management systems, digital asset management systems, and other repository platforms to manage all their different types of materials. The new products aim to simplify library operations through a more inclusive platform designed to handle all the different forms of content.”http://www.infotoday.com/cilmag/sep11/Breeding.shtml
• #15: We are trialling VuFind discovery and will be developing it to match user requirements at the University of London.This will definitely deal with our local bib data, and for us will have archives and ePrints included.Vufind and Blacklight are serious Open Source options for discovery – both are flexible and easy to work with, but VuFind is a better match to our local skillset and is especially interesting because Birkbeck, University of London are already using it live and they are an OLE project partner with us
• #17: Old web OPACs inherited concepts from the first and second-generation online catalogues from the 1970s and 80s, which were in turn heavily based on concepts from card catalogues.This system is what my systems librarian Simon Barron calls a ‘text adventure version of the LMS’.For a contemporary user base this really doesn’t make sense.I think,and argue that next-gen catalogues represent the first serious attempts to break these ties to the card catalogue still present in the WebPAC.I wanted to look at how readers understood the OPAC 2.0 system ahead of deployment to inform our approach
• #18: Quick summary – My view of online catalogues is heavily influenced by theory and practice from web usability.Web usability tends towards being cognitive constructivist in nature. That means they're based around an idea that our minds create understanding of external reality, and quoting from this reference here "within a systematic relationship to the external world”.We build ideas about how systems work in our heads, and apply these to systems.
• #19: There are strongly contrasting views of the catalogues. Some interviewees preferred the WebPAC, or some other OPAC, or nothing at all.I found catalogue user experience is very subjective. It was interesting the extent to which two people will rate Encore and the WebPAC as exact opposites when looking at the same things – for example he interface being being “clear” or “cluttered”.
• #21: Hidden slide not presented in my talk.
• #22: Hidden slide not presented in my talk.
• #23: What am I suggesting here: if you give the reader something that looks like Google then it encourages a style of search and behavior that reflect what we’d expect on a search engine.
• #24: It surprises me how much what we presented to the end user seemsto lead or influence behavior. That is, putting readers in front of a Google-style search interface tended to lead on to “Web like” behaviour.This was developed in an inductive way from the behaviour observed.We had not formulated this idea ahead of time then attempted to prove it.
• #25: The emotional experience of using the catalogue is an important part of the overall experience of our library.I found there was a strongly emotional or affective response to the catalogue beyond what you’d expect from a mere lookup tool.Much more than just about it being “nice to use” or “familiar”. A catalogue can be “a joy to use”. There is no reason for it to be a painful experience where you have to pay your dues and work through a painful process where you are judged deserving of a good outcome.Affect in discovery and search has become fashionable again recently and would be an interesting area for further study.
• #26: Blog post with more details on this.
• #27: Outcomes of staff focus groups are very interesting – there is a forward-looking aspirational view from library workers that we need to provide full discovery of everything in an easy to use way. This basic concept has been around since the 1960s at least, for example in JCR Licklider’s ‘procognitive machine’ (1965) but new technology and experience of current generations of search make this look like a pretty reasonable near-future goal rather than science fiction.‘We want the Moon on a stick’ is an actual quote from those focus groups. :-)Licklider, JCR (1965)Libraries of the future. Cambridge, MA: MIT Press
• #28: One approach doesn’t suit all readers. This was a really solid lesson learned from study and deployment.I always emphasize Encore as the best starting point but I don’t insist on it being used particularly by more experiencedreaders with more complex requirements and understanding of traditional OPACs. (Incidentally,we do still have that text adventure OPAC in use…)In 2012-13 (August to July academic year) Encore was 51% usage based on unique visitors versus 49% WebPAC usage.
• #30: It’s become a truism that as OPAC 2.0 and discovery makes better use of our bibliographic data they force us to sort out existing problems with our metadata. (These points I’ve made are expanded on in a blog post about a session I ran at a Library Camp unconference at Brunel University earlier this year.)Metadata optimization includes scoping reclassification and ‘tidying up’ legacy bibliographic data from previous systems migrations and integrations.What really attracts SMT is improvements in student experience and researcher experience. Needless to say bib data is not that engaging - so we present how improving metadata allows researchers to find things in our library and catalogue though and how it improves student experience for University of London users of our shared library services.
• #31: Counts from “facet filtering” helps to expose problems with metadata – look at an example here with ‘undetermined’ is 5,140 last I searched in our VuFind.That’s a valid coding for MARC but is hopeless for retrieval – there will be records in various languages in that tranche that are unfindable if someone trusts that ‘English’ or ‘Greek’ finds everything in that language, which is a completely reasonable expectation.
• #32: Here are some highlights from recent improvements to bib data at University of London.Further examples include:Invalid 006sBlank characters in the MARC leaderNo dates present in the 008041 fields with language codes run together (a legacy issue from UKMARC)We identified and updated about 105,000 problem codings in records already.