Metrics for Translation
Metrics for translation are the factors with which we can evaluate translation products and processes.
The growth of globalized and localized products has increased the need to manage translation costs. To reduce costs, organizations employ a variety of tools and methods to automate translation. These tools include software such as translation memory, machine translation, and content management systems. The tools are used in conjunction with methods such as writing for translation, topic-based writing, single sourcing, and collaborative development processes.
Each tool may provide metrics such as number of words and terms stored in translation memory, number of source and target languages, and accuracy of source-to-target term matches. Likewise, each method may also provide measurements such as word count reduction, development time improvements, and amount of content re-use. Specific metrics used to manage translation costs will vary by tool, method, product, organization, or even locale. The desired result is to minimize the overall cost of product globalization by managing translation costs.
Quality measurements for translation are harder to define. There is no standard definition for translation quality. As with source content, the quality of translated products and processes is subjective. Some metrics used for cost management can also be viewed as measurements of some aspect of translation quality, such as accurate term matches. However, differing cultural expectations and stylistic preferences can make standardization of metrics problematic. Ultimately, the usability of both source and target product content, evaluated separately, should be the final measurement of translation quality.
Different kinds of products, users and organizations require different metrics for translation to provide meaningful cost and quality evaluation. The challenge is to find the solutions that provide insight into ways of improving the overall quality and cost management of a globalized product.
Metrics for translation, translation quality, quality control, usability, translation costs, translation memory, machine-aided translation, language translation, technologies, 100% match, fuzzy match
Jeff Huset created this page to provide a description of the Comap node Metrics for Translation