Extraction and parsing of herbarium specimen data: Exploring the use of the Dublin core application profile framework

Herbaria around the world house millions of plant specimens; botanists and other researchers value these resources as ingredients in biodiversity research. Even when the specimen sheets are digitized and made available online, the critical information about the specimen stored on the sheet are not in a usable (i.e., machine-processible) form. This paper describes a current research and development project that is designing and testing high-throughput workflows that combine machine- and human-processes to extract and parse the specimen label data. The primary focus of the paper is the metadata needs for the workflow and the creation of the structured metadata records describing the plant specimen. In the project, we are exploring the use of the new Dublin Core Metadata Initiative framework for application profiles. First articulated as the Singapore Framework for Dublin Core Application Profiles in 2007, the use of this framework is in its infancy. The promises of this framework for maximum interoperability and for documenting the use of metadata for maximum reusability, and for supporting metadata applications that are in conformance with Web architectural principles provide the incentive to explore and add implementation experience regarding this new framework