Ten good reasons to use a translation memory
More than 20 years after the availability of the first commercial translation memory products, research indicates that while the vast majority of respondents use a translation memory system, less than 30% of translators use this type of tool for all projects of translation or even daily. Studies of translation memory usage among more technically advanced users show that the benefits of using a translation memory – besides reusing an existing translation – are still not well understood.
We discuss further below the main benefits that users of any commercial translation memory product can enjoy. In the author’s view, each of these benefits is quite convincing as a justification for using a translation memory tool, except the more exotic translation projects.
1. Reuse of systematic translation
This feature named this type of tool: the ability to store source and destination segments together, which allows a translation memory system to propose a translation in case the same source segment or a similar source segment appears in a future translation project.
Unlike traditional translation work, in which translators should actively search for previously translated sentences, translation memory systems automatically search for their memories and match the translator for re-use. In a translation memory environment, translators never translate the same sentence or a similar one twice.
Many commercial translation memory products not only support segment level matching, but allow translators to search for phrases/words level. Sub-sentential searches are useful for translators working with texts that are not sentence-level repetitive but have recurring phrases, placements, and idioms.
2. Automatic dictionary search
The two most relevant translation standards, ASTM F2575 and CEN EN 15038, set up project-specific multilingual terms as industry best practices. So the matter is: How do linguists access terminology data quickly and efficiently?
Keeping glossaries in Excel spreadsheets, word tables or PDF documents, make linguists stop the translation process and conduct an active search every time they come across an unfamiliar term. Besides, the result of this research is doubtful, since the translator does not usually know if the term unknown is actually included in the glossary or not.
On the other hand, a translation memory environment automatically highlights the terms in a segment that are available in the terminology management component. With automatic dictionary searching, translators simply select a target term and enter that term into the target segment when they find a term in a segment. Thus, having a comprehensive glossary allows translators to translate more terminologically challenging texts without interruption.
Example of a translation memory environment that illustrates text segmentation, formatting information filtering, and automatic dictionary search (available terms highlighted in red).
3. Easy integrity check
One of the most basic requirements for quality translation is that the translation is complete. Working in a traditional translation environment, there is no simple way to ensure that the entire text has been translated. Furthermore, in a translation memory tool an empty target segment stands out as a painful swelling, making it very difficult for translators, editors and proofreaders not to notice the translation lost.
4. Multiple file format support
All commercial translation memory tools support various file formats.
In other words: Translators can translate source documents that were created in an application that they do not know how to use and do not even have a license. For example, translators who use a translation memory system that includes a filter for Adobe InDesign can actually translate InDesign files without having Adobe InDesign installed on their computer.
This feature allows translators not only to work in a familiar interface, offering translation services for many file formats, but also to save thousands of dollars in licensing costs and training fees.
5. Simplified translation editing and proofreading
Both the main translation quality standards, ASTM F2575 and CEN EN 15038 require a person other than the translator to edit / revise each translation. In a traditional production environment, it is difficult to ensure that changes in a translation are transferred to similar translations in the future. There is no simple and effective way of alerting translators to the fact that they must translate a particular sentence in a certain way.
But even in environments where translation memories are used, there is evidence suggesting that editing / proofreading is often done outside translation memory tools.
There are two major problems associated with this practice. First, editing / reviewing a translated document in the target format (for example, Microsoft Word) or an intermediate format (for example, Adobe PDF) usually requires manual copy and paste, i.e. error prone, to update the translation memory with all editions made. This copy and paste between applications can take a long time, which leads to the second problem, some linguists neglect to update their translation memories after delivering the target document. This, in turn, results in translation updates where editors / reviewers repeatedly face the same mistakes that they have corrected in earlier versions of the document.
If, however, linguists use a translation memory not only for translation but also for editing / proofreading, they cannot only streamline the post-translation workflow by eliminating a number of non-value-added processes, they can also ensure that their translation memories are always up to date.
Representation of a traditional editing / revision process. Editing / revising in a translation memory tool can eliminate two or three non-value-added steps in this workflow.
6. More accurate estimate
In a business environment, it is important to make accurate forecasts about the time and budget required to complete a particular translation project. Working in a word processor environment, the only objective basis for any type of prediction is the total word count of the project.
In a competitive situation where multiple vendors applied for the same project, translators who base their estimates on the total number of words alone have a distinct disadvantage: Many longer documents in the technical scope have a significant percentage of internal repetition, which translation memory can identify, but the word processor cannot. Not to mention the scenarios where translation buyers send a translation memory together with the original document and expect a detailed quotation that reflects the advantages gained by the translation memory.
Even if there is no external pressure to provide the most competitive estimate possible, the analysis feature of a translation memory system allows translators to create more accurate project schedules. And better schedules allow translators to align more projects without increasing their stress levels.
7. Simplified collaborative advantage
When many translators work on an incremental update of a document, such as a user manual, the use of a translation memory is indispensable. However what happens with a project where only a few translators work on without any pre-existing translation memory?
Even in this scenario where many translators would argue that a traditional word processor is as useful as a translation memory tool, the author strongly favors translation memory.
Consider the following facts: Whenever a translation project involves multiple translators, stylistic and terminological consistency is a major challenge. Having a language-specific translation style guide in place can mitigate some of the style issues. However, exchanging, consolidating and sharing translation memories between translators allows all translators participating in a project to access inaccurate translations created by other members of the project. Collaborating using this imprecise translation technique ensures a greater degree of stylistic consistency than it would be possible in an isolated word processor environment.
Another great advantage of using a translation memory in a team environment is the ability to share project terminology in a very efficient way. Using the automatic dictionary search feature allows all translators involved in a project to use the same terms within and between documents. Sharing terminology in this way not only speeds up translation, but dramatically reduces the time needed to harmonize divergent terminology during the editing / review phase.
8. Improved quality assurance
Translating into a translation memory system provides translators with a simple and very effective solution to ensure the integrity of their translations, customer compliance, or project-specific glossaries, and stylistic conventions. There are different of cat tools, for instance, Wordfast and Trados QA Checker are examples of translation memories and ErrorSpy and QA Distiller are examples of dedicated translation quality assurance. They all provide extensive error-checking functionality.
All of these systems perform automatic comparative analyzes, i.e., they verify that a specific characteristic is present in the source segment and that its equivalent is also present in the target segment. These tools can be used to automatically identify terminology errors (design glossary deviations), translation inconsistencies (same source segment, different translations), number problems (values, formatting), etc.
Automated error checking, part of the quality assurance process, allows linguists to improve the formal quality level of translation in large projects much faster than it would be possible in a traditional translation environment which only depends on the manual check.
Linguists can use translation quality assurance tools such as ErrorSpy to generate automatic error reports in a batch mode or to manually adjust reports in an interactive mode.
9. Improve your translation productivity
As indicated in this author’s definition about a translation memory system, this type of tool automates many tasks related to translation. Examples of tasks that can be automated in a translation memory tool include automatic translation memory lookup, automatic match insertion, automatic dictionary searching, automatic integrity checking, automatic tag validation, and so on.
In addition, working on a translation memory tool has many ergonomic advantages, for instance, the source and destination clauses are visible in close proximity in a window, the source and target segments are linked and move in sync, the filters hide the distraction formatting so that the translators can focus on the translation task, etc.
All of these benefits are available from the first sentence that a translator translates into a translation memory tool. Thus, even in the most pessimistic scenario where there is never a repetition or a similar source sentence, translators should be able to translate faster because translation memories are simply better suited for translations.