Mach Translat (2006) 20:305–309 DOI 10.1007/s10590-007-9020-3 BOOK REVIEW Uwe Reinke, Translation Memories. Systeme – Konzepte – Linguistische Optimierung Saarbrücker Beiträge zur Sprach- und Translationswissenschaft. Europäischer Verlag der Wissenschaften, Frankfurt am Main, Peter Lang, 2004, 486 pp Dorothy Kenny Received: 2 February 2007 / Accepted: 2 February 2007 / Published online: 27 April 2007 © Springer Science+Business Media B.V. 2007 Translation memories (TMs) are now so commonplace in the translation profession, and training in the use of TM tools is such an integral part of most translator train- ing programmes, that the relative dearth of published research in the area remains something of an anomaly. True, several sources provide general introductions to the technology (Austermühl 2001; Bowker 2002; Somers 2003; to name just three), indi- vidual tools are often reviewed in professional journals (see also Corpas Pastor and Varela Salinas 2003), and a number of commentators have offered critiques of a tech- nology that is sometimes seen as undermining translator competence. But the evidence presented by critics often remains anecdotal, and there are only isolated instances in the literature of experimental research or attempts to provide general frameworks for those wishing to pursue research in the area (see for example, Bowker 2003, 2005). Compared to machine translation, research into the evaluation of TM tools seems particularly underdeveloped. Against this background, interested readers of German will welcome Uwe Reinke’s scholarly exploration of translation memories in a book that represents the culmination of several years of research in the field. Reinke’s motivation in conducting his research (at the University of the Saarland in Saarbrücken) was precisely the fact that translation memories seemed under- researched, as well as his constant observation that the matching algorithms in trans- lation memory systems very often produced hits that were completely at odds with what a human translator would consider ‘similar’ and hence potentially useful in the current translation context. His ensuing research thus leads him to tackle the question of similarity, as addressed in a variety of disciplines, and to explore ways of evaluating and improving the matching performance of TM tools. Throughout, Reinke keeps the human dimension in mind, seeing knowledge of human translation processes as a prerequisite for the development of computer-aided translation tools, and always D. Kenny (B ) School of Applied Language and Intercultural Studies, Dublin City University, Dublin 9, Ireland e-mail: dorothy.kenny@dcu.ie