Have you cleaned behind your glossaries?

Don't take this question too personally. After all, I'm not asking whether you've cleaned behind your ears, or behind your couch. But last week I asked the digital question, "Where do your glossaries live?" and this week I'm asking about the state of their hygiene.

One of my client-companies is quite proud (and justifiably so) of the considerable work they did a couple of years ago in building out a 600+ entry glossary in ten languages. They (or their language vendor, really) have hosted it on the Web, with read-only access to any translator who does work for them.

This model of glossary has the inestimable benefits of being universal, up-to-date and centralized - there is only one glossary - instead of being a patchwork of spreadsheets and tables on several different hard drives in several states of accuracy. It's set up for alpha-listed browsing and search, although the search function is not fuzzy unless you use wildcards, so some translators will not derive full benefit from it and may in fact miss terms.

While managing a sample translation for the client, I wanted to export the glossary to review it all at a glance, so I mentioned that. "Nope. That's not possible," the client told me, with more than a hint of pride. "We designed it so that there would be only one glossary in one format in one place. We don't want it exported or circulated unnecessarily."

Now, I'm in business to see my clients succeed, but that kind of mindset is just a tempting challenge to me, and as I managed the sample translation I deliberately looked for reasons why a hermetically sealed glossary like this was a bad idea. Naturally, I found one: The client had not cleaned very well behind their glossary.

Several industry-specific terms occur in the sample, and I knew the translators would be obliged to use the glossary. For instance, terms like "drive" occur in various combinations ("link drive," "offset drive," "drive mechanics," "rack-mount drive," etc.) in the glossary, and as I poked from one entry to another I noticed inconsistencies and contradictions in how "drive" was translated, notably in German. One entry gave "Laufwerk" as the translation, and another entry bore the note that "'Laufwerk' is obsolete."

The online model for hosting this glossary is a good one for several reasons, but it's not amenable to the healthy, periodic scrub that such databases should undergo. If the glossary were exportable, or at least visible in row-and-column format, these inconsistencies would be easier for translators to spot and address.

Interested in this topic? You might enjoy another article I've written called "Where do your glossaries live?"