DQF and DQF-MQM: Understanding the Relationship to MQM

How DQF-MQM fits

MQM Council documents the harmonization of TAUS DQF and MQM work, creating a DQF-MQM error typology. Its current history says that MQM Core has superseded that typology. This makes DQF-MQM important for interpreting existing scorecards, while requiring care before presenting it as the newest independent framework. MQM history.

Why teams still encounter the name

A team may receive old project instructions, reviewer spreadsheets or a tool configuration labelled DQF-MQM. The label alone cannot tell the recipient which categories were used or whether the scoring matches a current reference.

Our practical recommendation is to preserve the original configuration before attempting a migration. Identify the category names, severity definitions, weights and exclusions actually used. A renamed menu does not prove that old and new results are comparable.

An illustrative migration

Suppose a provider has six months of historical reports and wants to adopt a new scorecard. A useful exercise is to select an unchanged sample of translations and assess it under both configurations. Record cases where a finding moves category or changes severity.

The comparison can show which trends remain interpretable and which need a break in the reporting series. It is preferable to documenting a sudden “improvement” that is actually caused by different measurement settings.

This is an original migration example, not a TAUS implementation procedure or a claim about current integrations.

Connection to LISA and AI

The link to LISA is mediated by the development of quality models, especially MQM. It is not evidence that DQF is a renamed LISA product.

For an AI-assisted evaluator, our recommendation is to specify the actual category set and instructions instead of relying on a historical acronym in a prompt. Validate the mapping of the model's output to the intended scorecard before comparing results.