Why a specialized metric belongs in this collection
It illustrates an important design choice: a metric can be useful because it focuses on a defined type of content. That focus also limits the conclusions it supports.
A technical instruction and an advertising headline ask different things of a translation. If an assessment does not address a particular quality dimension, a favourable result cannot establish that the dimension has been satisfied.
An illustrative comparison
Consider two texts. One tells a technician which component to disconnect. The other introduces a vehicle to prospective buyers. Both need accurate language, but the second also relies on tone, persuasion and brand voice.
Our recommended exercise is to write down the intended use of each assessment before choosing a metric. Ask which failures the assessment should detect and which qualities need a separate review. This avoids selecting a familiar metric first and discovering its limits only after delivery.
The example does not reproduce SAE's categories or scoring rules. Those details require an authorized copy of the applicable document.
Relationship to LISA and MQM
SAE J2450 is a separate SAE publication, not a LISA standard. MQM's historical account lists it among the approaches examined during the development of a shared error typology. That makes it a useful point of comparison in the broader history of translation-quality evaluation. MQM historical account.
AI and scope discipline
An automated evaluator should not silently add its own preferences and report the combined result as the original metric. If a team wants to extend a scorecard for another domain, our recommendation is to describe the result as an adaptation and record the differences.
The presence of AI does not turn a specialized assessment into a complete review of every quality dimension. Automated suggestions also need to be checked against the content and instructions they are meant to evaluate.
Source record and related reading
SAE J2450_201608, including its stabilized status and stated style limitation. This article uses the public scope, not the full paid standard.
The historical statement drawn from MQM Council material is paraphrased under CC BY 4.0, with credit to the Council. See its licensing notice. No endorsement is implied.
Related: MQM, ISO 5060, LISA QA Model.