Metadata workflow based on reusing of original data

Abstract

The goal of the DiVA project is to develop a workflow where information from the original document created by the author can be reused to extract metadata for various purposes. Submission of item & creation of metadata There are two ways to submit an item and store the metadata in the system. 1. using a template (MS Word, Open Office, Star Office, LaTeX) 2. updating through the DiVA update. In both cases metadata are converted to XML in the DiVA XML:schema format and stored in the system. Dissemination Metadata can be generated from DiVA XML-Schema in many different formats: currently in Dublin Core(HTML-text and XML-RDF), TEI-Header, Endnote and Reference Manager, and MARC 21. All records are also exposed in unqualified Dublin Core to support harvesting by OAI-PMH service providers. Library systems A workflow has been created wherein the database record , originally created by the author, forms the basis for the bibliographical description of that record in the Swedish National Library system as well as the local system at Uppsala University Library. The record is harvested in MARC-XML format by the Royal Library in Stockholm and converted to MARC 21 "tape format" for import into the national library system and, after adding local information such as holdings, exported to the local system. Records can also be downloaded directly from the DiVA web site in MARC 21 "tape format" (ISO 2709), as well as MARC-XML format, for use in local library systems. Metadata for Uniform Resource Name National Bibliographic Number (URN:NBN) & Long-term Preservation Each item in the archive is assigned a unique, persistent National Bibliographic Number and Uniform Resource Name. After the URN:NBN is assigned, a record is created for harvesting by a resolver service. In addition to the delivery of metadata to the National Library, full-text files (currently PDF format with a DiVA XML-Schema file containing all archiving metadata) will also be delivered to support long-term preservation. A checksum will be computed independently for each document in both the Royal Library and DiVA archives.