In my introduction last week I briefly mentioned that I work as a "CAT" or localisation engineer. (Most of my career I've been a project manager, though.)
What does one do? As opposed to a localisation engineer at the "source" (software development), at the translation company side I don't work with the actual code. I rely on the material sent to us by our customers. So, I would get translatable files such as .properties, .resources, .resx or .po for applications or, more commonly, documents like Microsoft Word/PowerPoint/Excel or Adobe InDesign and work on them.
In a nutshell, I see my work as supporting project management and sales as a technical specialist.
At the beginning of a project (or for a quote)
The work usually starts with a customer wanting to know the costs of some content they need translated. We'll have to figure out the amount of content (translations are often paid per source word) and also if any extra work (not included in translation rate) is needed.
I would say the majority of the work is preparing various file types for translating in the translation tool. Even if you have a simple Word file, it may have translatable images or other embedded content that would need extracting before translation. If I come across uneditable images, I would check with the project manager if the customer wants us to prepare properly localised images or if they want to (limit costs and) just get the translations in, for example, a table format. If localisation is requested and my image editing skills are sufficient, I open the image in an editor, add text layers and make the original texts disappear. (If my skills are not sufficient, the images would be sent to other specialists.)
Office productivity tools also often seem to get all sorts of flashy new features that help with the creation of the document itself but might hinder translation, especially since the translation tool will have to keep up with the features. We in localisation engineering have to make sure all content is properly extracted and doesn't break during translation. (And then fix, if something manages to break.)
In software localisation files you usually have things like parametres or markup that I would check gets properly processed by the translation tool. Many file formats are quite standardised and, for example, the parametres would have little wiggle room in their appearance: %s, %d, {0}... But then you might have others with more descriptive names or custom formats: %PATH_TO_HEADER_IMAGE, [user]. I would check that these are recognised as special content because it affects costs, special content is protected in the tool and therefore more difficult to mess up during translation, and the tool's QA can spot any errors quite easily.
Localisation files can also have a custom structure where not everything is supposed to be translated so I'll have to tell the translation tool the specific bits that are edited.
Take for example a JSON file:
{ "menu": {
"id": "export",
"export": "Export",
"PDF": "As PDF",
"TXT": "As text file",
}}
You mustn't translate the value for id here but all the other content after a colon. You could either list all the properties that are translatable OR perhaps here you'd rather tell the tool what must not be translated.
One of my favourite tasks is to write regular expressions to specify the translatable or protected content in a file. 😄
The best approach to handle the files may vary from project to project.
At the end of a project
When the translation is ready and reviewed, the files may need some post-processing before delivery. I would check the layouts of the documents and if images are localised, I'll edit their layout as well and insert them in the document. For proper DTP (desktop publishing) work we have other specialists but if DTP is done by the customer, I'd just quickly check everything's in order.
Administration or maintenance
Sometimes I'm asked to do maintenance work on translation memories, or TMs (if you don't know what a translation memory is, don't worry! I'll tell you in a future post). As an example, we may get translation memory exports that need to be imported into our main TMs; there may be a document pair that should be aligned (source sentences are matched with translations, which is quite an easy job if you know the languages a little so you don't need a linguist do to that) and imported into a TM; or old file-based TMs need to be converted to server-based.
I can also help with user management: I'll create new user accounts when needed and give instructions on how the translators can connect the tool to our server, or give additional access rights. Sometimes translators have problems connecting and I'll help troubleshoot that.
When things don't work
Another common task in the job is troubleshooting. When a lot of people with different levels of technical prowess are using a tool in a hectic day-to-day work, you're bound to come across some interesting bugs and error messages (and sometimes the problem is between the keyboard and the chair, as they say). Sometimes I have helped with the same issue in the past so I know the steps to fix it, or a team member has documented the error and fix, but oftentimes the problem is all-new or the error message is very generic ("Object Reference Not Set to an Instance of an Object" 🙄). Then I'll usually check if I can replicate the issue and see how I can fix it or I'll search online in either the tool's knowledge base, which is quite good nowadays, or just any discussion forums. It's not uncommon that someone has already encountered the same issue!
What's that thing in the cover image?
Back in the day, the translation memory software Trados required a physical dongle. You'd pop this in the back of your computer and fire up the software. Sometimes if you forgot to take the dongle with you, someone would pinch it and then you'd be looking for a spare one the next morning! 😄 This dongle here is my honorary parting gift from one of my closest colleagues as I just left the company after 15 (!!!) years.
In closing
One of my favourite blogs on the topic is Paul Filkin's 'multifarious'. He works for SDL who is the developer of a widely used translation management tool, SDL Trados Studio. SDL's product documentation is dreadful, to be honest, so whenever I search for more in-depth information on a feature, I usually find the answer in Paul's blog.
Don't hesitate to ask if you'd like to hear more about something on this topic!
Top comments (1)
Very interesting topic! I'm also looking forward to hearing about TM's.