A Comprehensive and Standards-Aware Common Data Model (CDM) for Taxonomic Research

Abstract

The EDIT Common Data Model (CDM) (FUB, BGBM 2008) is the centrepiece of the EDIT Platform for Cybertaxonomy (FUB, BGBM 2011, Ciardelli et al. 2009). Building on modelling efforts reaching back to the 1990ies, it aims to combine existing standards relevant to the taxonomic domain (but often designed for data exchange) with requirements of modern taxonomic tools. Modelled in the Unified Modelling Language (UML) (Booch et al. 2005), it offers an object oriented view on the information domain managed by expert taxonomists that is implemented independent of the used operating system and database management system (DBMS). Being used in various national and international research projects with diverse foci over the past decade, the model evolved and became the common base of a variety of taxonomic projects, such as floras, faunas and checklists (see FUB, BGBM 2016 for a number of data portals created and made publicly available by different projects). The CDM is strictly oriented towards the needs of the taxonomic experts community. Where requirements are complex it tries to reflect them reasonably rather than introducing ambiguity or reduced functionality via (over-)simplification. Where simplification is possible it tries to stay or become simple. Simplification on the model level is achieved by implementing business rules via constraints rather than via typification and subclassing. Simplification on the user interface level is achieved by numerous options for customisation. Being used as a generic model for a variety of application types and use cases, it is adaptable and extendable by users and developers. It uses a combination of static and dynamic typification to allow both efficient handling of complex but well-defined data domains such as taxonomic classifications and nomenclature as well as less well-defined flexible domains like factual and descriptive data. Additionally it allows the creation of more than 30 types of user defined vocabularies such as those for taxonomic rank, nomenclatural status, name-to-name relationships, geographic area, presence status, etc. A strong focus is set on good scientific praxis by making the source of almost all data citable in detail and offering data lineage to trace data back to its roots. It is also easy to reflect multiple opinions in parallel, e.g. differing taxonomic concepts (Berendsohn 1995, Berendsohn & al., this session) or several descriptive treatments obtained from different regional floras or faunas. The CDM attempts to comprehensively cover the data used in the taxonomic domain - nomenclature, taxonomy (including concepts), taxon distribution data, descriptive data of all kinds, including morphological data referring to taxa and/or specimens, images and multimedia data of various kinds, and a complex system covering specimens and specimen derivatives down to DNA samples and sequences (Kilian et al. 2015, Stöver and Müller 2015) that mirrors the complexity of knowledge accumulation in the taxonomic research process. In the context of the EDIT Platform, several applications have been developed based on the CDM and the library that provides the API and web Service interfaces based on the CDM (see Kohlbecker & al. and Güntsch & al., this session). In some areas the CDM is still evolving - although the basic structures are present, questions of application development feed back into modelling decisions. However, a "no-shortcuts" approach to modelling has variously delayed application development in the past, but it now pays off: the Platform can rapidly adapt to changing requirements from different projects and taxonomic specialists.