Harvard University Library


<< Table of Contents

veritasHarvard University Library Notes, For Harvard Library Staff, Number 1325 May 2005

Harvard University Library Notes / May 2006 / No. 1331

New Virtual Collections Service from OIS

Virtual Collections (VC), a new system developed by the University Library's Office for Information Systems (OIS), enables Harvard librarians, archivists, and curators to create new, topic-based digital collections from data existing in Harvard catalogs, such as HOLLIS, VIA, and HGL. Using Virtual Collections, public users can now be guided to a curated collection selected from the millions of resources available through Harvard's union catalogs. Catalog records, including any links to associated digital objects, are harvested, loaded into the new virtual collection, and periodically refreshed through re-harvesting to capture any new cataloging changes in the source data. The Virtual Collections service includes a web-based user interface to provide browsing and searching capabilities that are confined to the individually defined virtual collection.

Harvesting records for a Virtual Collection is configured by individual collection curators and entails two major actions:

  • creating a harvest list, consisting of system/record ID pairs (HOLLIS/000000001, or VIA/olvwork00001, for instance), and
  • uploading that list for processing by the VC harvester.

The VC service utilizes a subset of rules established in the OAI-PMH (Open Archives Initiative Protocol for Metadata Harvesting) to request records. Records are transformed from their native forms (MARC, VIA, or FGDC) into a common metadata format (MODS, or Metadata Object Description Schema) shared by all collections, and loaded into the VC database.

A significant feature of the Virtual Collections service is the ability afforded collection curators to create and maintain controlled, collection-specific subject vocabularies, and to add these subjects to individual records. These local topics, called "categories" in VC, supplement but do not replace subjects (such as MARC 6xx fields) from the catalog of origin.

Because librarians, archivists, and curators wish to create collections that may range in size from a few dozen to a few thousand objects, the Virtual Collections service can be utilized in two distinct ways:

  • as a stand-alone application hosted by OIS, with modest customization of generic VC user interface possible, or
  • as an integrated, customizable application, in which search, browse, and record display functions are integrated into a collection's externally hosted web site.

Curators define the look and feel by selecting a banner and color scheme. The home page can present a narrative description of the collection. A navigation bar can provide links to related material. Because the Virtual Collection system uses an XML/XSL (eXtensible Markup Language/eXtensible Style sheet Language) model for web presentation, administrators of integrated collections can create their own XSL (and CSS) style sheets to customize the appearance not only of brief and full records, but of other screens as well.

The first use of Virtual Collections is a collection of Latin American pamphlets from the Harvard College Library (previously digitized through the Library Digital Initiative), viewable at http://nrs.harvard.edu/urn-3:hul.eresource:latampdc. The Open Collections Program will be using Virtual Collections as its primary database for future collections, including Immigration to the United States, 1789-1930, which is nearing completion now.

Future plans for Virtual Collections include:

  • making systems such as OASIS and TED harvestable for Virtual Collections, and
  • making Virtual Collections harvestable by outside institutions.

For more information on the Virtual Collections service, see the Virtual Collections page on the OIS web site at http://hul.harvard.edu/ois/systems/vc.


<< Table of Contents | << Previous Article | Next Article >>



Current Issue

Contact Library Notes


Return to the top.