In today’s globalized world, the need for translated content is more than human translators can cope with alone, which is why the interest in translation technologies is higher than ever. Translation memory (TM) is one of the results of this great interest in translation technologies. Translation memory software is a type of computer-aided translation (CAT) tool that stores previously translated phrases and sentences in a database and automatically matches and translates similar phrases in subsequent translations. This memory is extremely useful in ensuring translation consistency, increasing efficiency, and saving time and money. And these are not the only benefits for translators and translation services.
What is a Translation Memory?
Translation Memory, in its simplest definition, is a database that matches text segments from a source language to their equivalents in a target language and expands as new translations are made. These text segments, known as translation units, can range from a paragraph, sentence, phrase, or even just a sequence of words.
Thanks to Translation Memory software, translators do not have to start from scratch with each new translation project. Furthermore, businesses can rely on the work of other translators or previous translations even when they start working with different translators or translation agencies. After completing a translation project, the database of this memory expands with the reusable content from the newly made translations.
When was it Invented?
Translation Memory (TM) software, initially developed in the 1980s, was based on the concept of storing previously completed translations in a database and reusing similar expressions in subsequent translation projects. The primary goal was to enhance consistency and efficiency in the translation process. The first commercial of TM system was introduced to the market in 1989 under the name “Metatex.”
Since the introduction of the first TM software to the present day, there have been significant changes in translation technology, with many new features and capabilities added to these systems. Modern TM systems now come equipped with advanced matching algorithms, support for multiple file formats, and the ability to integrate with other Computer-Assisted Translation (CAT) tools, among other essential features.
How does Translation Memory Work?
Translation Memory (TM) software, which is a crucial tool for freelance translators, translation agencies, and businesses dealing with large-volume text translation, typically operates in a similar manner.
The TM software first divides a source text file into segments, often in the form of sentences or phrases. Each segmented unit is then compared to the segments stored in this memory’s database to check for matches. If a match is found, the relevant translation is automatically added to the target text.
This software also includes a rating system to determine the accuracy of matches. When a text segment in the source file is identical to the content in the TM, it’s considered a 100% perfect match. If certain parts of a translation unit match the database’s content, it’s referred to as a “fuzzy match.” In the case of a fuzzy match, the translator will need to make some modifications if they decide to reuse the TM content. If there is no match, the translator creates an entirely new translation for inclusion in the database for future projects.
The Benefits of Translation Memory
A translation memory fundamentally eliminates the need to repeatedly translate the same expressions. However, its benefits extend beyond this, offering many advantages for translators and translation agencies, such as:
- Consistency in Translations: Translators no longer need to repeatedly translate identical expressions. This memory ensures consistency in translation projects by handling this task. Moreover, it enables consistency across different translation projects related to the same topic, even if they are spaced out over time. Additionally, the need to research past translations for consistency purposes is eliminated.
- Savings from Repeated Translations: Every translation made in this memory software is stored in the database and can be reused at no additional cost. Therefore, projects with a high degree of repetition can be completed much more quickly and at a lower cost than initially perceived, leading to significant cost savings.
- High Quality: Project managers or translators can enhance the quality of this memory content to align with the company’s requirements, desired tone, and other specific preferences.
- Accessible Database: With each translation centralized in a database, it becomes accessible from any location, enabling different translators to work on the same project.
- Faster Translation Turnaround: Using TM allows for much quicker completion of a translation when numerous matches are available. This means more work can be completed in a shorter amount of time, more projects can be taken on, or the extra time can be devoted to producing original translations. Consequently, this positively impacts the earnings of translators.
When Should You Use Translation Memory?
When conducting regular translations, using a translation memory can yield excellent results, especially when there are repeated sections throughout the entire translation or in parts of it. One of the most significant advantages of a cloud-based TM is that translators can access the database via CAT tools from anywhere. In this context, the right time to use this memory is essentially at any time. However, using this memory for creative texts, literary translations, etc., may not be as beneficial because the number of repeated units will likely be minimal.
Nevertheless, this memory can be extremely useful for repetitive but different translations. It can provide significant time and cost savings for translations of technical documents and product manuals, legal and financial texts, video game localization, software localization, support portals, product descriptions, and e-commerce content.
How to Create a Quality Translation Memory?
To ensure translation quality while building a translation memory, here are some steps you can take:
- Work with a Reputable Translation Service Provider: Preventing low-quality translations from being included in your TM database upfront is much easier than trying to remove them later. Ensure that your database is built correctly and with quality.
- Provide Context to Translators: For the most accurate and high-quality translation, it’s essential that translators have a good understanding of the context of the text. Make sure you provide context that is compatible with the CAT tools used by translators.
- Prioritize Matches from the Database: Creating reliable translation memories and providing them to translators is essential to enable them to see and reuse appropriate translations immediately.
- Lock Segments with High Matches: Locking segments with high matches ensures that unwanted changes are not made to reliable translations that you’ve previously verified.
- Use Automated Quality Assurance Checks Before Importing Translations into TM: Even the best translators can make mistakes. Ensure that you perform automated quality assurance checks before importing translations into your TM. This step is essential for maintaining consistency in terminology and avoiding errors.
- Add Changes Made Outside of TM Manually: Any changes made outside of this memory should be manually added to the TM. As a result, your database will remain current.
- Provide Feedback to the Translation Service Provider: Share any errors, preferences, or changes you notice with your translation service provider. This ensures that the TM remains current while preserving quality and consistency.