Oversettelsessminne

Oversettelsessminne, often called translation memory (TM), is a database used in computer-assisted translation (CAT) that stores source-language text segments and their translated equivalents. Its primary purpose is to reuse translations in future work to improve consistency and efficiency.

A TM contains pairs of source and target segments, typically at sentence or phrase level, plus metadata such as project, domain, and creation date. When a translator works on new text, the CAT tool searches the memory for matches to each source segment. Exact matches can be reused immediately; fuzzy matches provide similar translations with a similarity score to guide the translator. Translations can be created or edited and then saved back into the memory.

Benefits of oversettelsessminne include greater consistency across documents and projects, faster turnaround times, reduced costs, and easier management of terminology and style across teams. It supports post-editing workflows by offering previously approved translations and can be integrated with terminology databases to support uniform terminology usage.

Limitations and cautions include the risk of propagating outdated or context-inappropriate translations if the memory is not properly curated. Quality depends on accurate segmentation, alignment, and reviewing of reused segments. Memory content may require licensing considerations if it contains proprietary material, and context may be lost when reusing isolated segments.

Relation to related tools and concepts: oversettelsessminne works alongside terminology management (term bases), quality assurance checks, and often integrates with machine translation workflows to propose matches or post-edits. It remains a core component of professional localization workflows, evolving with advances in CAT tools and MT.