Posts Tagged ‘search’

Find it @ Lincoln: looking forward to a new EBSCO discovery service in the Library

Posted on May 11th, 2012 by Paul Stainthorp

Following long, looong discussions, we have finally chosen a next-generation library discovery service for the University of Lincoln Library.

After reviewing the four major commercially-available discovery products (from EBSCO, Ex Libris, OCLC and Serials Solutions), and after making several reference visits to see the various products in action in UK university libraries…

(…drum roll…)

EDS logo…we decided upon, and have now bought access to, the EBSCO Discovery Service. Over the summer we’ll be configuring and testing the new system, and in September 2012 it’ll be launched as the new front-end search and discovery platform for the Library at the University of Lincoln.

This new service will provide a single point of search and discovery across nearly all of the Library’s collections, including our ‘traditional’ library catalogue, e-books & e-journals, the Lincoln Repository, archives & special collections, reading lists, and a wide range of specialist and general electronic databases. (N.B. it might not search all of these collections right from day one!) We hope that—along with some of the other new and improved services that are being introduced as part of the Library’s review of ICT systems—it will make it significantly easier and more straightforward to find and use the University’s library resources.

According to the SCONUL HE Library Technology wiki, the EBSCO Discovery Service is also used by:

We decided that EBSCO Discovery Service provided us with a familiar (yet flexible, powerful and ‘serious’) research interface, as well as a good fit with our existing and planned electronic database collections. We were also influenced by EBSCO’s plans to develop and integrate the A-to-Z e-journals knowledgebase and link resolver into the discovery environment.

We’ll be spending the next month or so configuring the system to search all of our collections, designing/branding the interface, training library staff, and working with other University departments on getting the most out of the new tools. We anticipate that early access to the system will be possible from the end of July onwards (though this is subject to change), with a ‘soft’ launch in time for student induction in September, and a formal launch/discovery party with free coffee for all, later in the year.

We have also decided that the service will be branded under the title “Find it @ Lincoln“. (Eagle-eyed readers will spot that this is the name we’ve been using for a while for our EBSCO LinkSource OpenURL link resolver.) Information about the new Find it @ Lincoln service, and about the project to develop and launch it at the University of Lincoln, will soon be available at: http://findit.library.lincoln.ac.uk/

I’d like to thank the staff of all four discovery software companies, for all the presentations, demonstrations & visits, for the information they made available to the University of Lincoln over the past few months about their products, and for the demonstrations and supporting materials they provided which were of such use in informing this first selection phase of our discovery project.

Many thanks also, to the several universities who received staff from Library for discovery-themed visits, and who patiently described their use of their own search tools and answered our many questions profound and otherwise.

Now watch this space :-)

More forgiving searches on the A-to-Z

Posted on April 13th, 2012 by Paul Stainthorp

EBSCO have improved the way the e-journals A-to-Z carries out journal title keyword searches.

Previously, the A-to-Z only matched on exact, ‘whole word’ searches; it was very unforgiving. Searching for science would not return results containing the word sciences. Stemmed / partial-word searching was (and still is) possible using an asterisk as a wildcard—e.g. scien* would return results containing science, sciences, scientific, scientist, etc.—but these kinds of search features don’t tend to be very popular with library users.

However, EBSCO have now introduced ‘stemming’ rules within the A-to-Z search engine.  This handles singular and plural forms such as science/sciences, and make for a more forgiving search. It also now allows searching using common journal abbreviations such as Br J Sports Med for the British Journal of Sports Medicine.

We used to have a small number of custom ‘redirect’ entries in the A-to-Z which picked up common misspellings of certain journal titles (for example, Journal of Forensic Science instead of the—correct—Journal of Forensic Sciences). These are no longer necessary and I’ve removed them from the A-to-Z.

Screenshot of the Journal of Forensic Sciences on the A-to-Z

If you need it, exact-title searching is still possible via the advanced search page, and you can still use an asterisk for partial-keyword stem searches.

List of cross-repository search tools

Posted on March 9th, 2012 by Paul Stainthorp

I’ve been wondering for a while why [national] aggregated/cross-repository search services haven’t really taken off – why aren’t they as well-known as union library catalogue services (e.g. Copac, which is part of the standard librarian’s armoury)?

Is it because aggregated search of repository-only content wouldn’t be particularly useful to researchers; perhaps because Google [Scholar] provides them with what they already need? Is it because no subset of all the repositories in the world would really meet researchers’ needs; i.e., they aren’t interested in finding articles just from one ‘showcase’, country-specific repo search tool? Because it’s too difficult? (Can’t believe that; not compared to the aggregation of catalogue data.) Or because OA is too far off 100% to make it a worthwhile exercise?

It’s certainly not for the want of initiatives and projects to build ‘em. A presentation at the recent UKCoRR members’ meeting made me realise just how many there are.

Here’s a list of ten eleven websites, tools and projects which relate to inter-repository search:

  1. Google Scholar (scholar.google.com), “a simple way to broadly search for scholarly literature” – the de facto cross-repository search tool. Google’s inclusion guidelines for webmasters (inc. of repositories). A journal article about finding repository content via Google (doi:10.1177/0961000606070587).
  2. Institutional Repository Search  (IRS) demonstrator from Mimas (irs.mimas.ac.uk/demonstrator), retrieves content “across 130 UK academic repositories”, from a project completed in 2009.
  3. KMi CORE (COnnecting REpositories) Portal (core.kmi.open.ac.uk/search), a newer project with its own project website and blog. “The CORE project aims to make it easier to navigate between relevant scientific papers stored in Open Access repositories. ” Recently extended by the ServiceCORE project
  4. OAIster (oaister.worldcat.org), developed by the library at the University of Michigan and adopted by OCLC in 2009. “More than 23 million records representing digital resources from more than 1,100 contributors.”
  5. OpenDOAR search (www.opendoar.org/search.php) – using Google’s Custom Search Engine (CSE) to search the full-text of material held in open access repositories listed in the OpenDOAR directory of repositories. At the time of writing this blog post, the service had been temporarily withdrawn since 25 January 2012.
  6. RepUK (repuk.ukoln.ac.uk), a project to build a central cache of metadata from institutional repositories in the UK (currently harvesting from 159 repositories).
  7. RIAN (rian.ie), a national portal to the contents of the institutional repositories of the seven university libraries in Ireland; “your route to Open Access Irish research publications” – this is the kind of thing I had in mind: why isn’t there one for the UK?
  8. ROAR (roar.eprints.org/content.html) – also uses Google’s Custom Search Engine across all 2000-odd repositories registered in ROAR.
  9. Subject and discipline-specific repositories including such venerable initiatives as arXiv (arxiv.org) and PubMed Central (www.ncbi.nlm.nih.gov/pmc): offering different approaches to aggregating content that—for the most part—ignore the role of the institution and work directly with authors and publishers, respectively.
  10. Mendeley (www.mendeley.com)… not searching repositories, but achieving much the same result, and, sez Les Carr, spanning the public/institutionalised (OA) and private/social (peer-to-peer) methods of providing access to papers.
  11. BASE (base.ub.uni-bielefeld.de/en/); “BASE is one of the world’s most voluminous search engines especially for academic open access web resources. BASE is operated by Bielefeld University Library.” (Added at the suggestion of John Murtagh, 12 April 2012)

Any others I’ve missed?

Now let us “thank” OAI-PMH (and quite possibly SWORD, too), for making all of this possible… other shared repository tools and projects include:  AEIOUJULIETNamesOA-RJORCIDOpen Depot, OpenDOAR, ORI, PIRUS2, RoMEO, and about 9,997½ more.

Ook Nog! Ook Nog! University of Liverpool student team win #DevXS library activity data prize

Posted on November 19th, 2011 by Paul Stainthorp

Four students from the University of Liverpool calling themselves Team Ook Nog took the prize for the best use of library activity data at last weekend’s DevXS student hackathon in Lincoln. Their application used the openly-licensed national OpenURL router data from EDINA and used it to build a search/recommendation tool for scholarly journal articles. You can see the fruits of their labour here

#DevXS - Team Boss Ook Nog

Jude-Thaddeus Ojiaku, Andrew Collins, Arnoud Pastink and Thomas Gorry built the Ook Nog site in a marathon development session over 30 hours in the Engine Shed. A simple Google-like search box (very Google-like!) displays results of articles and books derived solely from the OpenURL router data (example); each result has context-sensitive links out to dx.doi.org, OCLC firstsearch, CORE repository search, and Google Scholar. Clicking on any search result shows a chart of activity for that article, along with “See Also…” suggestions for other articles accessed by the same user in a similar timeframe. Take a look at the results.

From the DevXS wiki:

“Ook Nog is an interface for the data provided by openurl allowing you to search all of the data for any term and find search terms within their archive. By selecting any prior search term, you can then browse all search terms that were also performed by that user(s) within a small time period.

“All publications/searches are nodes. A node shares an edge with another node if a user has searched both nodes. We try to increase the chance of relevance by only showing neighbours of a node that were formed +- 90 days (a semester!).

“Despite no further tests of relevancy, the searches/publications found can be surprisingly similar (or amusing).”

The team from Liverpool pipped their traditional regional rivals to the library prize – Team MCR, made up of student developers from 3 different Manchester universities (University of Manchester, Manchester Metropolitan University, University of Salford). Team MCR built a working DevXS library app based around course reading lists with some interesting social ranking features, designed with great care using the Balsamiq wireframe UI tool, and making use of several open bibliographic datasets including the MOSAIC project data and Cambridge University Library’s search APIs. For their trouble, they picked up the #DevXS ‘social’ prize, awarded by the University of Lincoln Social Research Centre (LiSC).

DevXS was brilliant. Thanks again to Ian Snowley for the idea of donating a University of Lincoln Library prize. £250 in Amazon vouchers are on their way to Liverpool now.

Web history: Lincs to the Past

Posted on May 13th, 2011 by Paul Stainthorp

I’ve just returned from the formal launch event for Lincs to the Past, the new flagship website from Lincolnshire County Council which provides online access to “the cultural heritage collections” of the county.

Lincs to the Past launch party
The launch party was held at the Collection museum in Lincoln.

Lincs to the Past builds on a previous project of the county council called ‘Cultural Collections‘, which provided a unified search interface for resources held in various cultural-service catalogues (library, museums, archives).

The new website adds a whole load of interesting functionality on top of that single search, including:

  • Records collected together to form exhibitions
  • Very-high-quality digitised image browse (example) including rich navigation using Zoomify
  • User tagging and commenting
  • Faceted search (by date period, subject term, and domain: i.e. library, museum, or archives)
  • Online help

You can find Lincs to the Past at: www.lincstothepast.com

Searching the library through Facebook

Posted on March 31st, 2011 by Paul Stainthorp

I’d almost forgotten we’d done this: I wonder if anyone’s been using it?

http://www.facebook.com/universityoflincoln

University of Lincoln (UK) official page > Search the University library catalogue online

Screenshot of the library catalogue search in the University of Lincoln's official Facebook page

It’s just an HTML form that replicates a HiP general keyword search. I’ve done the same thing in our tab in Blackboard. We nicked the idea from (I think) Birkbeck, University of London.

Google magazines (slight return)

Posted on August 20th, 2010 by Paul Stainthorp

I’ve just recreated my list of magazines from Google Books for the University’s e-journals site.

Google now hosts 199 digitised magazine titles, and for the sake of 10 minutes’ work every few months it would be a shame to miss out on the extra full-text coverage, which often complements the “library” sources for a title.

E.g. for the frankly un-put-downable Estonian Journal of Archaeology (available as an Open Access (OA) journal from 2006-, and indexed in Art Full Text), Google provides the missing articles from 1997 (vol.1) up to 2006.

Record for the 'Estonian Journal of Archaeology' on the University of Lincoln's list of e-journals.

I’d like to be able to harvest the Google Books content to build my list using the standard mashlib toolkit (Google spreadsheets; Yahoo! Pipes; some coffee)… but while use of Google’s =ImportHtml() function is limited to 50 per spreadsheet, and because Google search pages block robots.txt files, I can’t figure out a way of doing so.

Instead, I’ve been copying-and-pasting the search results pages into an ordinary Microsoft Excel spreadsheet (thanks, again, Google, for making this possible through your magazine browse page), then using a custom Excel function to ‘unmask’ the URL hidden behind each hyperlinked magazine title.

Google Books magazine browse page, pasted into an Excel spreadsheet.

Finally, I use a bit of text-to-column splitting, search/replace, and filling-in of package-wide fields, to give me a compatible, tab-delimited text file which I then upload to our e-journals knowledge base (which happens to be EBSCO A-to-Z) – I used EBSCO’s custom notes feature to link to Google’s cover image to each entry in the file.