E.g., 11/18/2019
E.g., 11/18/2019

Mapping Today’s Translation Evaluation Practices and Needs

Assessing translation quality has been hampered by the lack of open and widely accepted tools and methods that support the automation required by today’s workflows, but how can this hurdle be overcome?

As part of a large-scale mapping of the translation and localization industries in order to push for a community-based solution to such hurdles, QTLaunchPad has been conducting a series of studies of, and inviting feedback from, stakeholders in these industries.

A recent snapshot of Europe (23 countries) consisting mostly of translators and LSPs looked at their practices and opinions on contemporary aspects of quality evaluation - QA. From this sample of almost 450 participants, we can see an interesting picture being painted:

How often is QA being carried out?

  • Nearly a third (31%) state that they perform quality checks in a systematic way for every job, while 16% report they never carry out such checks citing the relationship of trust they have established with their translators, who have been carefully selected and are highly experienced.
  • A further 15% carry out regular checks that are at fixed intervals or at random when circumstances allow.
  • And a minority (8%) perform one-off checks when there are changes made to the content.

How satisfied are freelancers and LSPs with their own approaches to QA?

  • A third (33%) consider their approach to QA to be sufficient to meet their needs.
  • Interestingly, 38% plan to improve upon them, but cost is cited as the greatest factor for not doing so.

As part of an on-going effort to map the practices and needs of the industry, QTLaunchPad is working in collaboration with GALA to conduct a large-scale survey, where stakeholders themselves can learn about current shortcomings and their potential solutions.

---

QTLaunchPad is a European Commission funded collaborative research initiative dedicated to overcoming quality barriers in machine and human translation and language technologies. We are a team of four world-leading research centers: the German Research Center for Artificial Intelligence, Dublin City University, University of Sheffield, and Athena Institute for Language and Speech Processing. Strong links with GALA, key players in the translation and localization industries, and translation communities around the world provide an excellent foundation for the development of the project’s work on translation quality assessment and estimation, mapping the translation and localization industry, and instigating an open community for a focused and intelligent push to overcome the current hurdles on the path towards high-quality translation and language technologies.

randomness