The Standard’s approach is based on the Multidimensional Quality Metrics (MQM) error typology. MQM comprises a hierarchical catalog of error types from which implementers can select subsets to evaluate translations of various content types. MQM is designed to support the analysis of errors in a wide range of translation-related scenarios.
These include the full spectrum of content types, comparing the quality of unedited machine versus human translation and the analysis of translation errors in localization environments. This practice emphasizes the “MQM Core,” a restricted subset suitable for most organizational applications and consistent with ISO 5060.