Translators’ requirements for translation technologiesuser study on translation tools

Resumo

This dissertation investigates the needs of professional translators regarding translation technologies with the aim of suggesting ways to improve these technologies from the users’ point of view. It mostly covers the topics of computer-assisted translation (CAT) tools, machine translation and terminology management. In particular, the work presented here examines three main questions: 1) what kind of tools do translators need to increase their productivity and income, 2) do existing translation tools satisfy translators’ needs, 3) how can translation tools be improved to cater to these needs. The dissertation is composed of nine previously published articles, which are included in the Appendix, while the methodology used and the results obtained in these studies are summarised in the main body of the dissertation. The task of identifying user needs was approached from three different perspectives: 1) eliciting translators’ needs by means of a user survey, 2) evaluation of existing CAT systems, and 3) analysis of the process of post-editing of ma- chine translation. The data from the user survey was analysed using quantitative and qualitative data analysis techniques. The post-editing process was studied through quantitative measures of time and technical effort, as well as through the qualitative study of the actual edits. The survey results demonstrated that the two crucial characteristics of CAT software were usability and functionality. It also helped to distinguish the features translators find most useful in their software, such as support for many different document formats, concordance search, autopropagation and autosuggest functions. Based on these preferences, an evaluation scheme for CAT software was developed. Various ways of improving CAT software usability and functionality were proposed, including making better use of textual corpora techniques and providing different versions of software with respect to the required level of functionality. Another major concern of the survey respondents was the quality of machine translation and its usefulness for creating draft translations for post- editing. In this direction, a part of this dissertation is dedicated to evaluation of machine translation, and investigation of the post-editing process. The findings of these studies showed which machine translation errors are easier to post-edit, which can be of practical use for improving the post-editing workflow.