Leveraging Translation Memories with Metadata

Few would argue with the observation that companies are increasingly waking up to the true value of their corporate assets and trying to make the best possible return on them. Translation assets are no exception. As translation memories (TMs) become increasingly valuable and shared and as TMs are shared in a tools-agnostic environment, TM metadata becomes all the more important.

This article was published in MultiLingual Computing magazine #113.

This case study explains and illustrates the importance of TM metadata on the conceptual level as well as its practical implementation. We look at the overall TM level as well as the level of individual segments, including but not limited to legal metadata; industry and content metadata (such as the TAUS Data Association ontology); metadata related to markup applied (LISA OSCAR TMX Level 1 or Level 2; other tools); metadata governing usage of placeholders; and LISA OSCAR Segmentation Rules eXchange (SRX) metadata.


  • Solution 1: Make order
  • Solution 2: Scope your project
  • Case A: Leverage across LSPs
  • Case B: Working on MT training sets
  • Cross-leverage on project level

In practice, numerous challenges to sharing TMs include the current deficiencies of the TMX 1.4b standard, restrictions and conditions from TDA about usage of shared TMs outside of TDA, and the unique challenges of intellectual property rights and restrictions from existing contractual or service-level agreements. All these can be addressed by sound TM management. This is a whole new discipline within the localization industry.

Know Thy Assets

Enterprise translation buyers and language service providers (LSPs) alike face this issue: It is difficult to know the value of legacy TM assets or even what is in them in terms of content, quality, relevance and so on.

As a project management officer, I am in charge of collecting TM assets for various purposes. Done manually across many tools and formats that we have been using on behalf of clients, this process has proven to be extremely time consuming, effort-intensive and painfully ineffective.

Adoption of Kilgray's TM repository has proven to help us know our clients' TM assets and create relevant asset subsets within hours rather than weeks and months.

Thanks to its extensive API, creation of project TMs across tools, product lines, content types and so on can be performed automatically within seconds to minutes as part of a well-engineered automated workflow that uses preset criteria. Let's look at the typical cases of how companies may manage their translation assets.

To continue reading this article, please complete this simple form below and download the full PDF version.

Fields marked with an asterisk (*) are required.

Get in touch

Got a question for us? We’d love to hear from you.  Have a challenge you’d like help with? We’ll give you a free assessment. Get in touch with us and we’ll get back to you right away.

Contact us > Call us >

Stay informed

Check out our blog for content on topics that matter to your business. Subscribe to get the latest.

Read our blog >