Abstract:

Biodiversity research needs a digital library on multi-source data. Biodiversity digital library, based on the virtual community, has similar features as digital libraries in data types, storage requirement and sharing methods. On the other hand, it has distinct features in terms of data mining and application. Based on an investigation of related digital library projects and cooperation with the Biodiversity Heritage Library and Internet Archive, we summarize the types of literature data in some kinds of digital libraries and briefly introduce Dublin Core and TaxonX standards which will be applied in the construction of the biodiversity digital library. Then, the present architecture of the biodiversity digital library, composed of data aggregation models and data processes, conversion and translation models and service models, is proposed in order to integrate multi-source data, construct the virtual community and provide specific data service to external web sites. Part of the implemented information system is demonstrated, and then some problems like copyright, OCR (optical character recognition) and the extension of massive data sets are discussed.

Key words: digital library, data standards, Dublin Core, TaxonX, full text search