Nomencurator: a nomenclatural history model to handle multiple taxonomic views

Evolutionary studies are generating increasing numbers of phylogenies which, in turn, sometimes result in changes to hierarchical organization and therefore changes in taxonomic nomenclature. A three-layered data model for a nomenclature database has been developed in order to elucidate the information structure in nomenclature and as a means to organize and manage a large, dynamic knowledge-base. In contrast to most other taxonomic databases, the model is publication-oriented rather than taxon-oriented and dynamic rather than static, in order to mimic the processes that taxonomists use naturally. The three-layered structure requires data integrity localized to each publication, instead of global data integrity, which relaxes constraints common to taxonomic databases and permits multiple taxonomic opinions: taxon names are made available as metadata within the model. Its prototype implementation, written in C ++, has an autonomous self-identification mechanism to avoid spurious data-inflation in a publication-oriented data model. Self-identification is also desirable for distributed implementations of the nomenclature database. Publication-oriented design also will make maintenance easier than for taxon-oriented databases, much of the maintenance workload being amenable to automation. The three-layered data model was designed for use by taxonomists, but is also able to provide concise, reduced expression for non-experts required in biodiversity research, for example.