A Short Note on Internationalization of Ontologies

Michael Conlon

TL;DR: A simple three step process for internationalizing ontologies is presented along with simple software to support the process

Abstract:

Ontologies contain text in the form of property and class labels, and annotations for helping ontology users determine what classes and properties represent. This text is best presented in a variety of languages to support use of the ontologies across the world and encourage their use for representing and sharing data. In this short note, a three step process is presented to enable a translator to provide text in any language for the text in an existing ontology. In the first step, text is extracted from the ontology in two languages -- a "from" language and a "to" language and placed in a spreadsheet for a translator to work in. The rows of the spreadsheet correspond to the classes and properties in the ontology. The columns correspond to the labels and annotations requiring translation. In the second step, the translator provides text in the "to" language for each label and annotation. In the third step, the spreadsheet is converted to OWL assertions, using the Open Biomedical Ontologies (OBO) robot tool, which can be merged to provide an updated ontology including the translations. A working example will be provided using the LANG ontology.

Software supporting these steps is available in an open source repository on GitHub.

Note to the organizers: This can be presented as a six minute lightning talk.

Keywords: ontologies, i18n, internationalization, open source