Schema driven assignment and implementation of life science identifiers (LSIDs).

Abstract

Life science identifier (LSID) is a global unique identifier standard intended to help rationalize the unique archival requirements of biological data. We describe LSID implementation architecture such that data managed by a relational database management system may be integrated with the LSID protocol as an add-on layer. The approach requires a database administrator (DBA) to specify an export schema detailing the content and structure of the archived data, and a mapping of the existing database to that schema. This specification can be expressed using SQL view syntax. In effect, we define a SQL-like language for implementing LSIDs. We describe the mapping of the view definition to an implementation as a set of databases triggers and a fixed runtime library. Thus a compiler for a domain-specific language could be written that would reduce the implementation of LSIDs to the task of writing SQL view-like definitions.