Linking libraries to the real world

Library search systems are always used in a specific context. Library customers use traditional
OPACs/catalogues and more modern discovery tools to find information about a specific
subject, for a specific project or course, etc.
There are two context related problems with using traditional catalogues:
Searches do not take into account the context and meaning of search terms, which results in a lot of ‘noise’
Results do not present information, but only descriptions (metadata) of materials that may or may not contain the information the customer needs, without possible relations between these materials 
The more modern library search and discovery tools offer some limited solutions:
Searches can be done in a number of catalogues and databases simultaneously, giving more results
Results can be narrowed down using explicit or implicit refining
Results not only describe physical materials (books, articles, etc.) but also digital content that can be retrieved online
FRBR is often used to link related editions etc. together
Recommender systems can suggest related material that other customers have shown an interest in 
But there are still some problems:
Search context: there is still no ‘intelligent’ determination of context/meaning of ambiguous search terms
Search context: still only databases are searched that the library provides access to
Results context: Physical and digital material still mostly only consists of content traditionally stored and provided by libraries: books, articles, basically texts published in a formal traditional procedure
Search/Results context: Lots of other content, data, metadata that is available on the web is not searched or found
We will not look at the search context problems, but we will focus on how to extend the results
context by using linked data. We will demonstrate this by presenting the “Dutch Culture Link”
proof of concept pilot project, a cooperation between the Library of the University of Amsterdam
and the Theatre Institute of The Netherlands, with support from the Digital Heritage Foundation
of The Netherlands.
In this project we are implementing both sides of the linked data concept: publishing and using
linked data. The Theatre Institute administers databases of theater performances metadata and
digital objects (sound, video, images, full text). These data are published as linked data.
The Library of the University of Amsterdam Aleph OPAC retrieves extra information from the
Theatre Institute performances linked data, based on the author and/or title in a bibliographic
search result, only if available.
The idea is that people searching for information about theatre related subjects in a library
catalogue will also be interested in related information about actual theatre performances.
We will talk about background, current implementation, choices, challenges, future plans and
developments:
Tools
Platforms
RDF
Vocabularies
URI patterns
Relationships 

Lukas Koster and Ad Aerts

Library search systems are always used in a specific context. Library customers use traditional OPACs/catalogues and more modern discovery tools to find information about a specific subject, for a specific project or course, etc.

There are two context related problems with using traditional catalogues:

  • Searches do not take into account the context and meaning of search terms, which results in a lot of ‘noise’
  • Results do not present information, but only descriptions (metadata) of materials that may or may not contain the information the customer needs, without possible relations between these materials 

The more modern library search and discovery tools offer some limited solutions:

  • Searches can be done in a number of catalogues and databases simultaneously, giving more results
  • Results can be narrowed down using explicit or implicit refining
  • Results not only describe physical materials (books, articles, etc.) but also digital content that can be retrieved online
  • FRBR is often used to link related editions etc. together
  • Recommender systems can suggest related material that other customers have shown an interest in 

But there are still some problems:

  • Search context: there is still no ‘intelligent’ determination of context/meaning of ambiguous search terms
  • Search context: still only databases are searched that the library provides access to
  • Results context: Physical and digital material still mostly only consists of content traditionally stored and provided by libraries: books, articles, basically texts published in a formal traditional procedure
  • Search/Results context: Lots of other content, data, metadata that is available on the web is not searched or found

We will not look at the search context problems, but we will focus on how to extend the results context by using linked data. We will demonstrate this by presenting the “Dutch Culture Link” proof of concept pilot project, a cooperation between the Library of the University of Amsterdam and the Theatre Institute of The Netherlands, with support from the Digital Heritage Foundation of The Netherlands.

In this project we are implementing both sides of the linked data concept: publishing and using linked data. The Theatre Institute administers databases of theater performances metadata and digital objects (sound, video, images, full text). These data are published as linked data. The Library of the University of Amsterdam Aleph OPAC retrieves extra information from the Theatre Institute performances linked data, based on the author and/or title in a bibliographic search result, only if available.

The idea is that people searching for information about theatre related subjects in a library catalogue will also be interested in related information about actual theatre performances.

We will talk about background, current implementation, choices, challenges, future plans and developments:

  • Tools
  • Platforms
  • RDF
  • Vocabularies
  • URI patterns
  • Relationships 

 

Attached files

elag2011_Koster (55,69 MB)


elag2011_Koster.m4v (42,41 MB)


Linking libraries to the real world (6,68 MB)


Editor: Milan Janíček
Last modified: 21.6. 2011 14:06  
Contact: +420 232 002 515, milan.janicek@techlib.cz