|11,106,374 NameBank Records|
NameBank is composed of a relational database stored in mySQL, an open source database management system. The figure shows a simplified form of entities contained in the database.
Sources and Agents: Data enters NameBank via Sources that are provided by Agents. A source could be an individual name, a list of names, or suggested changes to a NameBank record. An agent might is a person or agency that contributes the source. Multiple agents may be linked to a source and they may have different relationships to the source. For example, FishBase records vernacular names of fish and these are contributed to the Catalog of Life. We index the vernacular names from this source and both FishBase and the Catalog of Life Partnership are linked to the source as agents.
Agents are assigned editorial permissions that provide access to SOAP input methods for adding NameBank records and grouping them.
All NameBank record changes are journaled using mySQL 5 stored procedures that link record changes to agents and allow a record to be rolled back to any previous state.
Namebank records source taxonomic groups to provide organizational context to records. This is necessary because taxonomies often do not entrain all members of a taxon class. For example, we have indexed many global bird classifications most of which contain no or very little synonymy. The result is nearly 100,000 namebank records asserted to be birds but individual taxonomies that may only include half. In NameBank, these names are grouped in a pseudo-taxonomic group we call a Package. Source taxonomic groups are mapped to packages which draw upon the Catalog of Life to create a management taxonomy.
Distinct NameBank string concepts form the basis of a unique NameBank ID. A NameBank record ID is associated with a client class that qualifies the current review state of the record. A class ranges from 0-5. Low values refer to names that come from generalized non-authoritative sources or were located via algorithms like TaxonGrab or FindIT. Mid-range values refer to remote but authoritative sources. High values denote authoritiative review of the NameBank record.
Each NameBank record is assigned to a canonical group that is a normalized string index of the canonical name (name with no rank or author information). Misspelled names are in distinct canonical groups. Homonyms are in the same canonical group. Canonical groups are used primarily for search and retrieval of records and carry little informatic values.
Nominal groups unite lexical variants of the same name that including misspellings, variations in authorship, and names containing various nomenclatural or assertion annotations. A nominal group attributes qualifies the nature of the group item and a confidence level qualifies each entry.
Nomenclatural groups unite nominal groups based on homotypy. Each entry is annotated with a confidence attribute. Vernacular names are grouped with co-occurring scientific names to form vernacular groups or concepts.
Scientific name records may have a source Rank which is the default taxon rank but which may be overridden within ClassificationBank by a specific rank assertion. Vernacular names are qualified with a Language attribute.
NameBankIDs link records to Chresonyms. A chresonym references a name usage or assertion and is composed of a name and an object refernence that may be a citation or a physical specimen reference.
Mappings are NameBankIDs paired with remote name identifiers from other authoritative information sites. These provide the indices which can be exploited by LinkIT to conditionally link out to authoritative web sites. Mappings are grouped in Collections that contain additional URI metadata for composing a hyperlink.
uBio copyright © 2017 by The Marine Biological Laboratory