Cultural Heritage Collections: From Content Curation to Semantic Services and the Semantic Web (original) (raw)

Abstract

The action of collecting intentionally resources for a specific purpose, to organize them for personal use or for a particular audience, creates meaning. It has in itself a value which can be shared and used, just like descriptions or annotations to enable retrieval and manipulation of resources. All the actions of content creators, managers (e.g., librarians or museum curators) and users may be thought of through the concept of collection. Content creators often create a set of resources. Managers collect resources for a particular audience (e.g., the manuscripts of James Joyce or a collection of resources to support researchers in high energy physics). Users collect resources and organize them in their environment. Nevertheless, resources are most often described at item level and more rarely at collection level. The standards for the description of collections are not as stable and consistently used as standards for item level descriptions. As a result, while the work around resources is conditioned and driven by implicitly or explicitly created collections, those are often not represented in resource management systems. Recent advances in online services have emphasized interactions with users who can create their own collections and share them in Web 2.0 applications. Semantic representations of resources have also led to widening our conception of valuable resources because anything can be a resource of equal importance, a picture, a book, a city, an idea, and therefore also a collection. This article provides an overview of collection description practices, the integration of collections in different services, the metadata models for collection level descriptions, and the representations of collections on the Semantic Web. Different traditions: the computer science domain, museums, libraries, archives There are multiple definitions of a collection. The criteria used to aggregate resources in a collection are extremely different according to sectors and curatorial traditions. The traditional interpretation of library collections is associated with tangibility, ownership, a user community and a service (Lee, 2000). The Conspectus methodology was even created by the Research Libraries Group in the U.S. to assess the strengths and weaknesses of research libraries collections and therefore engage libraries in developing complementary collections on a regional or national basis, for instance. Andy Powell (1998) has described the diversity of traditions for collection definition: almost always, the collections of 'archives' delineate themselves: they relate normally to a specific person or institution. The collections of 'libraries', on the other hand, should be de-lineated by the purpose for which the library exists: by the information needs of their user populations. In contrast, the collections of 'museums', areagain-delineated somewhere between those two extremes. They can perhaps best be conceived as a bridge between the collecting desires and interests of specific people or institutions; and the information needs-in the widest sense-of those who might use the resulting collection.

Figures (3)

MICHAEL portal at European level and the IMLS Digital Collections and Content IMLS DCC in the US" gather together descriptions of digitized collections of heritage resources. Both hav been used as a basis for the harmonization of item level resource descriptions in the digita environments and the creation of access services at both item level and collection level. These registries use a definition of collections provided by the metadata creators. As a result collections are very heterogeneous. Metadata creators use the traditions of museums, libraries archives and others. In addition, collections are defined at various levels of granularity. A 100-iten collection is represented next to a 50.000-item collection. Michael Heaney (2000) describec information landscapes as the organization and access to information at various degrees of oranularity and specialization, through the definition of collections. Attempts to mix different levels of granularity in the same portal raise major issues, such as the constraints created by vatious uset expectations on what they will find on a given portal Nevertheless, the lack of consideration for collections also creates challenges, for instance because the item level descriptions in certain collections are too similar and create biases in the digital library system (Foulonneau, 2007). The possibility to collect both item level description: and collection level descriptions can however create opportunities, to compensate for incomplet records or add context to decontextualized records fetched in a third party aggregation such as at OAI-based portal (Foulonneau et al., 2005).

MICHAEL portal at European level and the IMLS Digital Collections and Content IMLS DCC in the US" gather together descriptions of digitized collections of heritage resources. Both hav been used as a basis for the harmonization of item level resource descriptions in the digita environments and the creation of access services at both item level and collection level. These registries use a definition of collections provided by the metadata creators. As a result collections are very heterogeneous. Metadata creators use the traditions of museums, libraries archives and others. In addition, collections are defined at various levels of granularity. A 100-iten collection is represented next to a 50.000-item collection. Michael Heaney (2000) describec information landscapes as the organization and access to information at various degrees of oranularity and specialization, through the definition of collections. Attempts to mix different levels of granularity in the same portal raise major issues, such as the constraints created by vatious uset expectations on what they will find on a given portal Nevertheless, the lack of consideration for collections also creates challenges, for instance because the item level descriptions in certain collections are too similar and create biases in the digital library system (Foulonneau, 2007). The possibility to collect both item level description: and collection level descriptions can however create opportunities, to compensate for incomplet records or add context to decontextualized records fetched in a third party aggregation such as at OAI-based portal (Foulonneau et al., 2005).

In order to support the description of heterogeneous collections, registries need to create a col- lection description model for all types of digital cultural content, whether from libraries, archives, museums or archaeological sites, for instance. Certain metadata models, such as MARC, were mainly conceived for item level descriptions but also used for collection level descriptions. MODS guidelines” specify a particular use of certain elements in the context of a collection level description, e.g., extent’. VRA-Cote” for visual resources explicitly defines a model for collection level descriptions in addition to item level and work level descriptions. EAD”, which is used in the archival domain, includes both collection and item level descriptions, as well as information on the organization of different collections. The Research Support Libraries Programme’ in the UK designed a simplified collection descrip- tion format for libraries. It then inspired the work conducted for the Dublin Core Collections application profile”, and thereafter the IMLS DCC collection description format and the metadata format used for the MICHAEL collection registry (Figure 2). Collections were even integrated into the CIDOC-CRM model, ie., the Conceptual Reference Model for cultural

In order to support the description of heterogeneous collections, registries need to create a col- lection description model for all types of digital cultural content, whether from libraries, archives, museums or archaeological sites, for instance. Certain metadata models, such as MARC, were mainly conceived for item level descriptions but also used for collection level descriptions. MODS guidelines” specify a particular use of certain elements in the context of a collection level description, e.g., extent’. VRA-Cote” for visual resources explicitly defines a model for collection level descriptions in addition to item level and work level descriptions. EAD”, which is used in the archival domain, includes both collection and item level descriptions, as well as information on the organization of different collections. The Research Support Libraries Programme’ in the UK designed a simplified collection descrip- tion format for libraries. It then inspired the work conducted for the Dublin Core Collections application profile”, and thereafter the IMLS DCC collection description format and the metadata format used for the MICHAEL collection registry (Figure 2). Collections were even integrated into the CIDOC-CRM model, ie., the Conceptual Reference Model for cultural

petspectives. several Cultural Institutions nave started using Puckr and simuar sites to pubust their resources. Users can create their own collections. Moreover, tools for the publication of data on the Semantic Web ate growing rapidly. This allow: envisioning that different actors can create and publish collections, while only publishing a "map' of that collection (e.g., in OAI-ORE) and its description, without collecting the content itself The Dublin Core Collection application profile or a derivative implementation can be used fo: describing the collection content. VoID can also be used for the description of the metadata set published on the Semantic Web. Figure 3 illustrates the complementarity between cor vocabularies to support services for cultural heritage items and collections in digital library systems as well as on the Semantic Web. Nevertheless, the co-existence of multiple levels ot granularity and of extremely heterogeneous types of resources remains a challenge for which new visualization tools and smart services designed for large datasets provide new answers.

petspectives. several Cultural Institutions nave started using Puckr and simuar sites to pubust their resources. Users can create their own collections. Moreover, tools for the publication of data on the Semantic Web ate growing rapidly. This allow: envisioning that different actors can create and publish collections, while only publishing a "map' of that collection (e.g., in OAI-ORE) and its description, without collecting the content itself The Dublin Core Collection application profile or a derivative implementation can be used fo: describing the collection content. VoID can also be used for the description of the metadata set published on the Semantic Web. Figure 3 illustrates the complementarity between cor vocabularies to support services for cultural heritage items and collections in digital library systems as well as on the Semantic Web. Nevertheless, the co-existence of multiple levels ot granularity and of extremely heterogeneous types of resources remains a challenge for which new visualization tools and smart services designed for large datasets provide new answers.

Loading...

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.

References (15)

  1. Beagrie Neil. Plenty of Room at the Bottom? Personal Digital Libraries and Collections. D-Lib Magazine, 2005 vol. 11 no 6.
  2. Boardman Richard, Sasse Angela M. Stuff Goes Into the Computer and Doesn't Come Out: a Cross-tool Study of Personal Information Management. In: CHI '04: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. New York: ACM Press, 2004, p. 583-590.
  3. Lagoze Carl, Fielding David. Defining Collections in Distributed Digital Libraries. D-Lib 1998 vol. 4 no 11. [Viewed 10 March, 2011]
  4. Entlich Richard. FAQ: Is There a Good, Comprehensive Catalog of Web-accessible Digitized Collections Available on the Internet? RLG DigiNews 2000 vol. 4, no.6.
  5. Foulonneau Muriel, Arms Caroline, Shreeves Sarah L. Sharing Resources by Collection: OAI Sets and Set Descriptions. Digital Library Federation Spring Forum, Austin, TX 2006. [Viewed 10 March, 2011] Foulonneau Muriel. Report Analysing Existing Content. Minerva Project. [viewed 10 March, 2011]
  6. Foulonneau Muriel et al. Using Collection Descriptions to Enhance an Aggregation of Harvested Item-level Metadata. JCDL '05: Proceedings of the 5th ACM/IEEE-CS joint conference on Digital libraries. 2005 p. 3241.
  7. Foulonneau Muriel. Information Redundancy Across Metadata Collections. Information Processing & Management. 2007 vol. 43 no 3 p. 740-751.
  8. Heaney Michael. An Analytical Model of Collections and their Catalogues. Bath: UKOLN, 2000. [viewed 10 March 2011]
  9. Lee Hur-Li. What is a Collection? Journal of the American Society for Information Science, 2000 vol. 51 no 12, p. 1106-1113.
  10. Lourdi Irene, Papatheodorou Christos, Doerr Martin. Semantic Integration of Collection Description: Combining CIDOC/CRM and Dublin Core Collections Application Profile. D-Lib 2009 vol. 15 no 7-8. [viewed 10 March 2011]
  11. Powell Andy. Collection Level Description -a Review of Existing Practice. eLib, 1998. [viewed 10 March 2011]
  12. Renear Allen H. et al. Collection/Item Metadata Relationships. Berlin: Humboldt University, 2008 p. 80-89 [viewed 10 March 2011]
  13. Rousseaux Francis, Bonardi Alain. Benefiting from Piaget to improve our Collections Browsing Tools? IADIS Applied Computing 2007 Salamanca, 18-20 February 2007. [viewed 10 March 2011] Rousseaux Francis, Bonardi Alain. Parcourir et constituer nos collections numériques. CIDE 2008 [viewed 10 March 2011]
  14. Stvilia Besiki, Jörgensen Corinne. User-generated Collection-level Metadata in an On-line Photo-sharing System. Library & Information Science Research. 2009, vol. 31 no 1, p. 54-65.
  15. Urban Richard J, Twidale Michael B, Adamczyk Piotr. Designing and Developing a Collections Dashboard. In: Jennifer Trant and David Bearman (eds). Museums and the Web 2010: Proceedings. Toronto: Archives & Museum Informatics. 2010. [viewed 25 March 2011]