Bibliographic records such as those in the International Union of Crystallography (IUCr) dataset often contain author names and other entries originating in non-roman scripts, but listed in transliterated form. Bibliographic tools should be capable of smoothly handling datasets with multiple representations of item data, including data in non-roman scripts. Because transliteration results in data loss, legacy data sets cannot be refactored by automated means. However, with an ever closer connection between bibliographic data sets and authoring tools (Endnote, Zotero, Mendeley), it will become possible to crowdsource this work, by opening a path for the reuse of data sets which are enriched by end users during the authoring process. The first step toward such an objective would be to extend an existing reference manager to work smoothly with multilingual data, for RDF import, editing, and RDF export. submitted 22 Jan '11, 05:26 fbennett |