Mike Dillinger, Principal of TOPs and past president of the Association for Machine Translation in the Americas, discusses how small-scale, in-house MT systems let you test the usefulness of MT for your projects and develop the know-how to use it better than your competitors, without the risk of large investments. MT is a viable option for language service providers who want to decrease costs, accelerate delivery time, and show that they understand technology. The key is that you have to know how to choose it and how to use it. Free MT systems from Google and Microsoft are inexpensive and easy to access, but they have very significant risks and limitations. Large-scale enterprise MT systems offer excellent tools and performance, but are far too expensive and complex for most LSPs.
Mike Dillinger, PhD is Principal of TOPs (Translation Optimization Partners) Consulting, an independent consulting group that helps clients optimize communication in global markets. He focuses on aligning authoring with business goals and optimizing machine translation at clients including Apple, eBay, HP, Welocalize, Cisco, Borland, Raritan, Motorola, and others.