Ecoinformatics Conference Service, International Conference on Ecological Informatics 6

Strategies for Resolving Natural Language Differences in ILTER Data and Metadata

David E. Blankman, Xuebing Guo, Honglin He, Jianhui Li, Chau-Chin Lin, Sheng-Shan Lu, Chih-Jen Burke Ko, Aikiko Ogawa, Eamonn O'Tuama, Herbert Schentz, Wen Su, Kristin Vanderbilt, Bert van der Werf

Last modified: 2008-09-13

Abstract


The International Long Term Ecological Research Network (ILTER) strategic plan calls for widespread data exchange among member networks to support broad scale synthetic studies of ecological systems. However, natural language differences are common among ILTER networks and seriously inhibit the exchange and interpretation and proper use of ecological data. We review five potential approaches to translating metadata and data among ILTER sites to resolve this problem: 1) provide one central language plus a native language; 2) translate to all languages represented by member networks; 3) translate to a subset of all languages represented by member networks; 4) provide the native language only; 5) leave the decision on what to do to the ILTER member. We conclude that, although there are no ideal solutions because not all ILTER providers have sufficient resources, ILTER data providers should still bear the burden of translating at least a minimal set of metadata elements to English. In addition, providers with sufficient resources should strive to provide more complete metadata translations into English, and, if possible, into other languages. ILTER should also invest resources to create multilingual ontologies for highly structured data and metadata that could be used in future automated translation systems.