Translation Memory Copyright
Sharing Translation Memory Data Aligned from Third-Party Documents
TTMEM accesses existing third-party documents in multiple languages, segments the documents, and matches the segments in the different languages creating a bilingual database of translation memories.
Short phrases and slogans are not copyrightable in themselves because individually they do not provide sufficient originality to be protected by copyright. Since the distributed database is comprised of short phrases, each such phrase is not copyrightable because it is insufficiently original and protectable under the copyright laws. The translation memory database is provided for the purpose of generating a translation of other, independently generated text. Therefore, the translation memory database, along with any translations generated with it, are not likely to ever be substantially similar to any protectable expression in the source documents. As a result, the translation memory database and its intended use will not constitute or result in any infringement of copyright in any source documents. The entire purpose is to generate an instructional, factual translation memory database that will not provide any meaningful amount of protectable expression in the database or through the use of the database. The purpose is to generate a translation memory database that does not supersede the object of the original creation. Instead, they provide a new work with a completely different purpose and character.
Source Texts
The source documents are factual and technical. There is no threat to any market, for the original works in question or protectable derivatives of them, which would result from the proposed use of these documents. Most, if not all, of the source documents will be instruction materials that are downloadable from the Web. Consequently, the proposed use and resulting internal document copy and databases will not harm any market or reduce any revenue to the copyright owner for such source documents. This factor supports the fairness of the proposed uses of the source documents.

Some translation memories may contain names and detailed company data. Since these data come from documents which are easily downloadable from the Web, they are not copyright protected.
The process of producing translation memories by staff:
›› Does not infringe any copyright since the original text cannot be recreated AND/OR
›› Does not violate the principles of fair use since the purpose of the translation databases is completely different from the original document AND/OR
›› Does not violate any copyright laws, since the data come from documents which are easily downloadable from the Web.