Your questions answered: Terminology management and SDL MultiTerm
Fourth webinar summary is here! Robert Muirhead was the moderator for the last webinar by SDL Trados, and the speaker was Tom Imhof from localix.bix. After three amazing webinars, it was hard to keep up with the others, but he made a great presentation that turned out to be one of my favorites. The information presented by Tom was extremely useful. A good refresher’s course for all of us, whether you use Trados or not!
Here is the link to the full video: https://sdl.webex.com/sdl/lsr.php?RCID=5c8a1d42845b4e49ab3c0f318b8f3006
I am aware that my recent posts summarizing these webinars have been way too long, but I truly believe that they all contain valuable information, particularly for beginners, as a great introduction to terminology work. In this webinar, Tom covered basic theory up until minute 18, and then he moved to a practical example by creating a termbase with MultiTerm 2015. So this post will only cover the theory and then you can watch the rest starting on minute 18.
Tom started by defining a termbase as a “database that stores and manages terms with their translations” and holds unlimited amount of languages. It is usually integrated with a translation memory system. Since data is structured it can be reused by other programs such as machine translation systems.
A termbase entry is concept-oriented, that is, one entry contains the lexical representation of a concept in multiple languages. So, two words that have the same writing (homonyms) have to be registered in two separate entries. For example the word “bank” in German means “bank” and “bench”, so they have to be registered separately, as compared to a general dictionary in which the two meanings are registered under one entry.
He went over the theory of the semiotic triangle according to which we have an idea of an object in our heads about, for example, a tree. However if I see a pine tree, I would have that concept of tree in my head, but someone who lives in another region might think of a different type of tree, so the two concepts are different and therefore need to be recorded separately. The entry may also include metadata such as the source and a context in which the term is used. The term should appear in nominative singular form, but if you want to add the plural form you can add it in the entry as a data field. If using two regional variants, such as English US and UK, particularly in a translation environment, you should be consistent and use the same language variant in your termbase but may add a data field for the other variant. Following these rules will provide better results when you are using a translation memory system.
The data that you may store in a termbase is varied. Here are some examples he gave: Brand- or industry-specific terms, acronyms and abbreviations, slogans, preferred terms, new term candidates, confirmed terms, frequently-used terms, forbidden or outdated terms, terms that should not be translated, such as product or brand names.
The structure of a termbase entry includes three levels: (1) entry level (metadata that refers to the entry, such as subject, notes, images, status, etc.), (2) language –or index level– (metadata that refers to each specific language), and (3) term level (data that refers to the specific term, such as gender, context, definition, etc.).
The language or index field is the most important one because it contains the actual terms in each language. As a fun fact, the reason why the language level is also called index level is because from a database point of view the language field was the only one that was indexed and was easy to search in the very first databases that existed (such as the first versions of MultiTerm), but nowadays everything is indexed but the “index” term still remains to refer to the language level.
So besides the index field, we also now have descriptive fields that contain metadata that describe the context in which a term is used, status (an indication if the term is new, preferred, outdated, forbidded, etc.) and other descriptive information such as images and sound files. You may also enter free text in a descriptive field or attach multimedia files or picklists. We also have system fields that are created automatically with each entry and include information such as the username and date/time.
The rationale behind the question “Why should we do terminology work” is that there is a large number of term variants in use throughout an organization and terminology work helps us improve communication within a company and with the outside world (e.g., clients) that ultimately will result in major cost savings. From a marketing perspective, terminology work allows us to create corporate language identity and avoids liability issues by making sure that terminology, for example in machinery or user manuals, is consistent and accurate and will prevent users from making mistakes. Terminology work saves time and money, especially if it is done before a project starts so that all stakeholders “speak the same language”.
As I said at the beginning of this post, after reading this information you may skip to minute 18 and watch the video through minute 56 when the Q&A session starts. So the Q&A session goes from minute 56 to 1:27 which I also recommend you to watch.
I hope you have enjoyed this series of webinar summaries as much as I have. Thank you again to SDL Trados for offering such great webinars and help us understand terminology better.
Also, remember that I am saving these webinar summaries in my blog’s cloud under “training” and “videos”, for future reference.