Computer-assisted translation (CAT) tools have become an essential part of a translator’s toolkit, as well as a standard for Language Service Providers (LSPs). Though many translators prefer to do things the ‘old-fashioned way’, CAT tools can be a useful resource for increasing speed and turnaround time as well as increasing quality assurance. This is especially important for LSPs with a large (and large number of) projects.
This portfolio serves to demonstrate the skills I have acquired over the past semester in my Translation Technology course at the Middlebury Institute of International Studies with professor Adam Wooten.
LSP Simulation
Vega Language Solutions
After spending the first half of the semester learning the fundamentals of SDL Trados as well as practicing client interaction, we were tasked with carrying out a simulation of a mini in-house translation team localizing a project.
- This first involved locating an adequate source text, which, in a real-life scenario would be sent to us by the client, not the other way around. in our simulation, the organization, Anti-Slavery International asked us to translate a word document of the text from their Who We Are page into Latin-American Spanish and Brazilian Portuguese.
- In order to identify potential translation issues, it was necessary to run a pseudotranslation. This identified an issue of extracting text embedded in an image, to which the solution involves changing a filter on Trados to ‘Extract document properties for translation: All document properties’
- We then prepared a Statement of Work (SOW) in preparation for our ‘kick-off’ meeting with the client. This document included an overview of the project, including project steps, costs, and timelines.
- The next step was to prepare the project by setting up a translation memory (TM), a termbase (TB), a glossary, and a style guide in order to maintain consistency across translations. TM was created during project set-up in SDL Trados. TB was set up first
- Once the project setup was complete, we assigned the roles of project manager (PM), translators, and reviewers.
- The entire Translation, Editing, Proofreading (TEP) process was carried out in SDL Trados
- Final verifications were carried out by the reviewers, including content and structural checks.
- Following quality control, the final deliverables, including the source file, the pseudotranslation, the Latin-American Spanish translation, the Brazilian Portuguese translation, the assets (TM, TB, glossary, and style guide) were ready for submission.
The entire process was a valuable experience that allowed us to put what we learned throughout the semester into context. However, it did not come without challenges that were necessary to overcome. Below you can see a short presentation on these challenges and the lessons we learned from them:
Regex
SDL Trados has a number of customizable settings that can be used to assure quality in translation, such as checking for formatting errors, mismatches across translations (punctuation, number format, etc.), but obviously, it cannot control for every possible circumstance. For these types of case-specific instances, we can make use of a very useful function known as Regular Expressions, or Regex, for short. This semester we were able to learn how to structure and utilize these expressions, even creating our own regex functions that would be useful for our language pairs.
Fixing Quotation Format (commas outside of quotation marks)
As an example, let us consider formatting of ending punctuation and quotation marks. In English, punctuation is generally left inside the quotation marks, but in Spanish it is generally left outside the quotation marks.
Under normal circumstances, one is limited to simply searching for commas throughout the entire text, but this could be very time-consuming, considering the size of the project. However, using regex, we can tell the search function to only look for commas or periods at the end of a string inside quotation marks. Additionally, we can tell it to move these commas outside of the ending quotation mark so that we do not have to go in and manually fix them. Here is an example taken from regex101.com, a useful site for testing regular expressions:
This is easily implemented in SDL Trados by using the Find and Replace function and selecting Regular Expressions.
Fluent
The ultimate goal of translation technology is not to replace human translators, contrary to what many people believe or even fear, but rather to make the entire translation/localization process easier and more efficient, by removing tedious aspects and roadblocks.
One particular organization we were introduced to this semester exemplifies this goal wholeheartedly: Fluent.
The aim of this organization is to facilitate localization for languages other than English, due to the fact that the paradigms on which software localization has been based are largely English-centered. That is, “translations that map one-to-one to the English copy.” As most other-language speakers know, this paradigm is doomed to fail, as there are particularities of each language that do not directly correlate with English.
Project Fluent gives localizers the ability to address translation issues on a case-by-case basis, through a user-friendly platform where the localizer can program the system to meet a language-specific need.
After reading the Syntax Guide, we were able to use the Playground on projectfluent.org to come up with our own custom syntax for language-specific functions. I decided to apply this to create a function that is sensitive to gender identity, including non-binary users, in Spanish.
This is simply a case in which someone is added as a “new friend” on a social media platform, and in Spanish both the adjective and noun must reflect the gender of the subject. The masculine form is considered the “default,”, and though it does not necessarily imply biological gender, it can still lead to social repercussions. Although an officially-recognized form is yet to exist, a variety of gender-neutral forms are in common use to replace the masculine o or the feminine a, such as ‘@’ from the example below and similarly ‘x’. Regardless of what the localizer chooses to use, a gender-neutral option can indeed be accommodated.
I believe that this kind of project speaks to the true aims of translation technology development and its open-source nature will allow it to continuously grow and serve users.
In conclusion, translation technology is not here to replace us, but serves as a valuable resource that should be embraced in the pursuit of quality translations.
“Words travel worlds. Translators do the driving.” – Anna Rusconi
