Title: "Crowd Computing: Making the Crowd Valuable for Industry"
"For over ten years CA Technologies has been developing and using machine translation technologies and tools to support software localization activities. Like most large software vendors, CA Technologies continuously improves its processes to reduce the cost of translation and increase the number of languages it supports. The most expensive and time-consuming phase of the localization process is the post-editing by human translators of the output produced by automatic machine translation. Our hypothesis centers on the development of semi-automatic management platforms that allow the integration of crowd computing for reducing the cost of post-editing phase of software localization. In this type of systems, quality of the translated result hinges on the excellence of the work done by the users in the crowd. Trusting the individuals in a social network and their capacity to carry out the different tasks assigned to them becomes essential. In addition to ensuring the high quality of the translated text, we also must ask ourselves whether the process makes sense in an industrial context, given confidentiality and cost considerations. And in general, we are looking to define the characteristics of an industrial process that make it suitable for crowdsourcing."