| Subcribe via RSS 

Cat out of the bag: Here Comes the Sun

November 21st, 2006 | by mm

From the The Tool Kit by Jost Zetzsche.

Here is a find that you don’t make every day: unbeknown to much of the translation industry, Sun has released very large and very good glossaries for many of their products at https://g11nportal.sun.com/sungloss. You can freely register and then use the online glossary tool to look up individual terms or term lists, or you can also request downloadable .csv files that you can then use in a tool of your choice. It’s really quite excellent and a MUST for anyone who does anything even remotely connected to software translation.


While we’re on the subject, here is a little afterthought on how to structure glossaries or termbases.

Aside from the source and target term, there is obviously a great variety of other information that could and should be in a termbase. This includes context, subject area, client, project, synonyms, grammatical indicators, and many other fields. In my workshop for project managers last weekend, I stressed the importance of entering as much information for each term or phrase pair as you think will be useful later on. However, be as stringent with your time while entering additional information as you can be. Unless your particular tool makes use of information about the gender of a term, there is probably no reason to enter it (unless it is a neologism where the gender may still be in flux). After all, you obviously are expected to know that as a professional translator. If your tool has a hard time showing great amounts of context information, it does not make sense to be too verbose. The rule of thumb is that the less time you spend per entry while still entering everything valuable, the more likely you will continue to do it. And “do it” you should. Some of you have heard me stressing ad nauseum that the area where the translation industry is linguistically most sorely lacking is terminology management. And I think this is true for all stakeholders in this industry.

Yesterday was a crazy day for me. I was busy with all kinds of things all day, and I wasn’t able to start my actual project, a 4000-word translation for an SAP implementation, until late afternoon. There were virtually no matches in my otherwise very handy SAP translation memory, but almost every segment had numerous matches from my SAP termbase. Everyone who has ever translated SAP stuff knows how awkward its terminology is, but it really doesn’t matter how awkward it is if your termbase automatically displays all matches for each segment. I (almost) got done last night (and even made it to my 6:00 meeting – albeit at 9:00).

Here is another thought: depending on your language combination, you will want to adjust your strategy of how you enter terms or phrases into your termbase. For instance, if you work from an agglutinative language like Turkish or Finnish, it certainly does not make sense to add only one form of a verb or noun as the source because chances are that it will rarely find a match. Translators from English, on the other hand, may be able to be much more sparing with their entries of source terms. Of course, even in a language with little flexion it may make sense to enter not only “computer” but also “computers,” and not just “crash” but also “crashes” and “crashed.” In fact, “computer crash” or even “the computer crash” would also be good candidates.

And lastly, it is important to realize how your tool allows you to reuse records from the termbase. If it allows you to enter records with a keyboard shortcut or a mouse click, you will have to enter terms differently than if it were only a lookup tool. For instance, take the record

English: Computer
German: Computer, Rechner

Several target entries for one source entry per record is fine as a reference, but not if you use your tool interactively during translation. In that case, your records should look like this:

English: Computer
German: Computer

English: Computer
German: Rechner

That way you can automatically enter one term or the other.

Leave a Reply