Interview with Prof. Dr. Klaus-Dirk Schmitz

Prof. Dr. Klaus-Dirk SchmitzKlaus-Dirk Schmitz is Full Professor of Terminology Studies and Managing Director of the Institute for Information Management at Cologne University of Applied Sciences in Germany. The formal training of Prof. Schmitz was actually in Computers and Linguistics. He holds a diploma in Computer Science and Mathematics and a Ph.D. in Applied Linguistics and Information Science, both from the University of Saarbrücken. His teaching and research activities focus on terminology theory and terminology management as wells as on software localization and computer tools for translators. Numerous articles and books where he acts as author, co-author and editor show his excellent expertise in these fields (see list of publications). Prof. Schmitz has also taken a leadership role in national and international forums e.g. as former president of the Council for German Language Terminology (RaDT), former president of the International Information Centre for Terminology (Infoterm), chairman of the German Standards Committee on Systems for managing terminology, knowledge and content, vice-president of the German Terminology Association (DTT), vice-president of the International Terminology Network (TermNet), and member of several advisory boards. He has participated in several national and international projects related to terminology and language technologies (TDCnet, SALT, Euro TermBank, German Terminology Portal etc.) and developed a terminology eLearning portal (ELCAT). And as he always keeps up with the time he is now one of the five partners in the innovative project “Terminology as a Service” (TaaS) meant to establish a sustainable cloud-based platform that provides core terminology services.

1. You are very active in the terminology world could you tell me a bit about your activities?

Klaus-Dirk Schmitz: Besides teaching, academic administration, writing articles and doing research, I’m very much involved in professional associations and strategic commissions that deal with terminology and related fields. Currently I’m contributing to several working groups of the German Terminology Association (DTT) for updating our Best Practice Folder for terminology work, and I’m involved in the planning and conceptual organization of several congresses, such as our CHAT track (Creation, Harmonization and Application of Terminology Resources) during the tekom TCworld conference in Wiesbaden in November 2013, our next DTT Symposion in Mannheim at the end of March 2014 with papers on copyright, return on investment and resources in the field of terminology, or the next TKE Congress (Terminology and Knowledge Engineering) in Berlin in July 2014. In the field of terminology standardization, I’m currently working on finalizing a German version of ISO 26162 “Design, implementation and maintenance of terminology management systems” and we have every two weeks a web meeting on the revision of ISO 30042 (TBX).

2. You have also coordinated a lot of projects on terminology and language technology. One of the most relevant current projects is called TaaS. Could you please give some information on the evolution of the TaaS project and what is your role in the project? And which other projects do you coordinate?

Klaus-Dirk Schmitz The aim of the TaaS project[*] is to help language workers as well as CAT and MT tools by creating a web service for terminology, combining access to existing high-quality resources such as IATE and EuroTermBank with methods for finding new terms where they are not yet in the databases but might be findable on the web. TaaS also provides an infrastructure for sharing individual and company terminology resources, clean-up of raw term candidates, and for the validation of new terms and their translations. The pre-final version of the TaaS platform was presented at the very successful project’s mid-term review in Luxembourg. Our University was mainly responsible for the user needs survey, the functional specifications and the data modeling during the first year of the project, and we will concentrate on the usage scenarios, the testing and the evaluation during the second year. At the moment, we are only involved in one other research project called TIPPS (Terminology Information Policy, Portal and Service). TIPPS is a 3-year Turkish-German bilateral project with two main project partners, Hacettepe University in Ankara and Cologne University of Applied Sciences.

3. The Cologne University of Applied Sciences offers a master in terminology and language technology can you tell us about it?

Klaus-Dirk Schmitz: Our Institute for Translation and Multilingual Communication is a CIUTI member and has a very long tradition in academic education of translators and interpreters. After the Bologna process started, we were the first German University in that field moving from a Diploma to Bachelor and Master degrees. Our BA is called “Multilingual Communication” with two foreign languages (English, French, and Spanish) and three areas of specialization (business, culture and media, translation). The language professionals are educated on the MA level: We offer a MA in Specialized Translations (part of the EMT network), in Conference Interpreting, and in Terminology and Language Technology. The MA Terminology and Language Technology is unique in Germany; in four semesters students learn everything about theory, practice and tools in the area of terminology and language technology, but also about technical writing, software localization and specialized translation (only between German and English and with a concentration on IT texts).

4. Could you think about some approaches on cooperation between EU institutions and universities in the field of terminology? What can be done to enhance collaboration between the EU institutions and academia?

Klaus-Dirk Schmitz: Besides the internship opportunities, I see two main areas of cooperation between EU institutions and academia. One is related to projects, the other to theses. We have two project courses in our MA curriculum, a terminology project and a language technology project. We always look for application-oriented tasks (from real life outside the academia) that we can implement and conduct in these project courses. As an example from last semester, we converted with a group of 12 students several bilingual lexicographical printed glossaries to a concept-oriented termbase from which data for a smartphone app can be generated. Another very successful way of cooperation with industry can also be applied to the EU bodies. Students select a research topic for their master thesis in collaboration with an industrial partner, very often as a result of discussion during an internship. The university, the company and the students benefit from such types of collaboration. Similar application-oriented projects as well as coordinated topics for master theses can also be conducted together with EU institutions.

5. You are in permanently contact with language technology developers. Which are the new trends for terminology managing systems?

Klaus-Dirk Schmitz: Most of the terminology management systems on the market are highly developed and offer almost every feature a user will need. They are also very flexible so that they can be adjusted to specific organisational environments and needs. What could be improved is the support of the terminology workflow and the different status an entry, a concept or a term will have during the process of elaboration and verification. We also see the trend to correlate terminology management much more to the Web, i.e. web-based terminology management tools, using the web as a resource for terminological retrieval and involving the crowd into terminological activities. As a consequence, terminological data have to be more interoperable with other language and content resources, such as ontologies or other Linked Open Data resources. And the last trend that I have observed is that – similar to translation memories – many terminological resources have grown during the last decades and together with a lot of reliable data, a lot of inconsistent and impure data are stored. Quality assurance and cleaning features are urgently needed as a functionality of terminology management systems or as additional tools.

6. Prof. Schmitz what role does terminology play in German scientific and industrial world?

Klaus-Dirk Schmitz: Germany, or better, the German speaking area in Europe has a long tradition in terminology activities, both in the scientific and the industrial world. Eugen Wüster as the founder of terminology science and a lot of terminological literature published in German are indicators for this. The first huge main frame term banks (TEAM and LEXIS) were established in Germany and many of the software tools for terminology management were developed in the German speaking area. In terminology standardization as well as in the field of teaching and training terminology, this region plays a pioneering and still leading role. Professional associations for terminology and related fields are very active in Germany: The German Terminology Association (DTT – www.dttev.org) with about 300 members is very active, and the German TechWriters Association (tekom – www.tekom.de) with more than 8.000 members always implements a terminology track as part of its huge conference on technical communication (2012: 3.700 participants). Also industry in Germany pays more and more attention to terminology management, not only to support translation, localization and technical writing, but also as part of its corporate language activities and means of knowledge organization.

7. How important are international standards for terminology exchange and how should developers and standardisation bodies work together in this respect?

Klaus-Dirk Schmitz: There is no doubt, that interchanging terminological data between different users, system and application requires standards. If there are no standards available, individual conversion routines have to be implemented in every single case, and the owner of the source data has to be consulted to explain the meaning of each single type of information. That’s the reason why standards bodies such as ISO have started the development of an exchange standard for terminological data already in the beginning of the 1980ies. Currently we are happy to have ISO 30042 (2008) that specifies TBX (TermBase eXchange) as an XML-based exchange format for terminology. Some of the tools developers have implemented import and export routines that support TBX or the a little bit simpler version TBX-Basic. Currently we are working on an updated version of TBX and on some amendments to the standard that will facilitate the understanding and implementation of TBX features. But there is one huge dilemma that hampers the wide usage of TBX: Modern terminology management systems are very flexible and allow the user to create all the data categories that he or she needs, what implies that only the user knows the semantics of the own specific data categories, and not the developer of the TBX interface. But usually the user is not able to implement TBX import and export routines, due to the complexity of TBX. But we are working to solve this dilemma.

8. You are also an expert in the localisation field, which role plays terminology in software localisation?

Klaus-Dirk Schmitz: Terminology is very crucial for a lot of products and documents, but especially for software products. There are several reasons for this. Software very often introduces new concepts for new features of the program, and there exist no established terms for these new concepts in the target language; terminologists have to coin new terms for these new concepts. Software tools consist of many part: the software itself with menus, dialog boxes, error messages, etc, but also printed documentation, online-help, installation guides, tutorial, sample files and so on. Under a hard time pressure for bringing the localized versions at the same time to at least the important markets (simship), many people are involved in the localization of one program. This requires the use of consistent terminology through all parts of the software. And last but not least, terminology is the means of communication between the program and the user. If the terminology is not transparent, not appropriate and not consistent the user will be frustrated and will not be able to use the software properly.

9. We all know translators have not always the time and resources to do terminology work, which is the best approach for translators to be able to do an acceptable terminology work?

Klaus-Dirk SchmitzEven if translators are very often under time pressure, they should do terminology work following the concept-oriented approach. Otherwise the terminological data will be unsystematic and therefore unusable over the time. Concept orientation does not mean that translators have to elaborate concept systems and have to supply every terminological entry with a definition; it requires (only) that all terminological information belonging to one concept should also be managed in one terminological entry. Synonyms should be stored in the same entry, homonyms in different entries. And with a good terminology management system and a good preparation of terminology work, translator can easily create entries that contain more than two terms. Before a translation job starts, an entry template can be filled with default values for the subject field, the client, the term status, and if not done automatically with the name of the editor and the date. During the translation process, it is only necessary to add the source language and the target language term to create a valid terminological entry.

10. IATE is the EU’s Interinstitutional Terminology Database with entries in 27 languages. What do you think about it?

Klaus-Dirk Schmitz: I have used the free version of Eurodicautom, one of the EU termbanks that was the major source for IATE, already in 1985. Long before Windows and the Web exist in the way we know it today, I connected via modem to a host mainframe computer and sent a question for a term, and after a short time, I got the answer. The idea to include all EU resources into one system is really great and helpful for everybody. And there have been a lot of activities to improve the quality of the terminological information, what I really appreciate. If people are criticising some entries in IATE, I tell them that you always have to cross check terminological information you find, especially in the web. And if you have this in mind, IATE is a huge and in most cases very reliable source of information for terminologists, translators, interpreters and technical writers.

11. What is your vision about introducing the representation of concept system in a terminological database?

Klaus-Dirk Schmitz: Single concept relations are easy to maintain in a terminology management system; you need just some data categories such as /superordinate concept/ or even simpler something like /see/ with a link to another concept and may-be an attribute for the type of relation (e.g. partitive). Such concept relations are also easy to elaborate. But elaborating a whole concept system for a domain or sub-domain is really a hard work and very time consuming. In opposition to academia, nobody has time to develop a concept system in real working environments for translators, techwriters and terminologists. Currently there are only few applications that allow to represent or to maintain concept systems: Webterm (www.iim.fh-koeln.de/webterm) displays concept systems elaborated by students in the framework of terminological theses, and i-term (www.iterm.dk) together with i-model is a terminology management systems together with a graphical concept modelling tool. And there is coming a really new approach (www.coreon.com) to combine terminology management and taxonomy management, implementing something what we discuss since many year under the slogan terminology and knowledge engineering.

12. TermCoord tested some extraction tools to find the appropriate one to extract terms for proactive terminology projects. In your opinion which is the best approach for terminology extraction?

Klaus-Dirk Schmitz: The best approach fir term extraction is the human approach, but of course this is too slow and too expensive. If you want to use software for term extraction, tools with linguistic knowledge are much better than tools that operate on the basis of statistical and heuristic methods, especially for multi-word terms and discontinuous terms. The disadvantage of linguistic term extraction tools is only, that they are not available for all languages; only the major languages are supported. Some tools developers combine the linguistic and the statistical approach. In our TaaS project (see question 2) we try to implement linguistic based term extraction tools at least for all European languages in order to provide users with the necessary services for their terminology work.


[*] The research within the project TaaS has received funding from the European Union Seventh Framework Programme (FP7/2007-2013), grant agreement no 296312.


Interviewer: Matilda Soare

matilda-soareMatilda Soare graduated in Translation and Terminology (German/English) from the University of Bucharest and got a MA in Terminology and Language Technology from the Cologne University of Applied Sciences with a thesis on “Terminology of Terminology Science”. During her MA studies Matilda worked as a research assistant in projects dealing with student oriented and project based studies in the university. Matilda is passionate about terminology, project management and language technology.