Looking for a description of translation specifications? You can find it at this page.

Tranquality stands for “translation quality”. Tranquality provides an introduction to understanding translation quality and how to evaluate it using the Multidimensional Quality Metrics (MQM) framework.

MQM, developed by the EU-funded QTLaunchPad project, is a framework for describing a family of translation quality metrics. MQM can be used by anyone, free of charge, and customized translation quality metrics can be defined within the framework.  However, the framework itself cannot be modified and still called MQM.  This is to avoid multiple incompatible versions.

The translation error typology in DQF from TAUS is subset of the typology in MQM, so you subscribe to DQF, you are using MQM. Several translation tool vendors have implemented either full MQM or the DQF subset.

Translation quality management isn’t just about finding and fixing errors using metrics. It is about considering quality from the very beginning to prevent errors before they even happen. At the heart of this process is reliance on translation specifications, as defined in the ASTM F2575:2014 standard. Specifications allow translation requesters and providers to agree in advance upon expectations. When appropriate specifications are provided to all stakeholders, everyone knows what is expected, and project results can be evaluated in terms of those specifications (and metrics based on them), rather than being evaluated on the basis of various personal opinions. This promotes efficiency and satisfaction. But the quality management perspective also means that you take a proactive approach to fixing problems before they occur and anticipating the needs of stakeholders so that you are creating the right content for them in the first place.

MQM Definition

The official MQM definition is published by the German Research Center for Artificial Intelligence (DFKI) in association with the QT21 project. For convenience, a copy is also available here.