Identitas: Semantics-free and human-readable identifiers

Abstract

In this article, we present a new approach to identifiers, one that aims to improve the management of ontologies. This approach overcomes some of the main flaws associated with the existing approaches, providing alternative solutions. Ontology identifiers are the key for each entity defined in an ontology, and enable a unique and persistent reference to each term. The form of identifiers has been the subject of discussion, which has resulted in a number of different schemes. It is often recommended that identifiers for ontology terms should be semantics-free or meaningless. One practice is to use numeric identifiers, starting at one and working upwards. However, this has a number of disadvantages: it does not allow for concurrent development, is relatively hard to read, and it is difficult to detect errors when an identifier is misused. From the perspective of development, solving these issues could facilitate the process of building and managing ontologies. Here, we suggest random identifiers to enable concurrent development, while exploiting the proquint library to overcome the problems of memorability and pronounceability. Finally, a checksum is implemented to prevent the occurrence of errors while accessing relatively similar identifiers.

Availability and Implementation:

The software is available from https://github.com/Nizal-Shammry/identitas-j. It has been integrated into environments for ontology development such as Tawny-OWL and Protégé.