Skip to content
Platforma LivoLink » Blog » Znaczenie odległości edycyjnej przy tłumaczeniach

Editing distance in translations – what exactly is it?

    Tags:

    During translation, every character matters – it affects meaning and comprehension. Modern translation tools now rely on editing distance to automate translation processes. But what exactly is it, and why is it so important for translators?

    Editing distance in translations – what is it?

    Editing distance is a mathematical measure of how many operations are needed to transform one string of characters into another. This allows you to analyse how different two words are. Each operation – adding, deleting, or replacing a character – counts as one step.

    The simplest form of editing distance is Hamming distance, developed by American mathematician Richard Hamming. It only applies to strings of the same length, e.g., cat and kat, lama and rama, or mama and tama. In all cases, transforming one word into another requires changing only one letter, giving a Hamming distance of 1.

    In modern translation technology, more advanced measures are used, such as Levenshtein distance.

    What is Levenshtein distance?

    Levenshtein distance was introduced in 1965 by Vladimir Levenshtein, a Russian mathematician and researcher influential in computational linguistics.

    This widely used variant of editing distance calculates the minimum number of editing operations required to transform one string into another. The Levenshtein algorithm considers three main operations:

    • Insertion – adding a character (e.g., changing plan to plans requires one insertion, distance = 1)
    • Deletion – removing a character (e.g., changing trees to tree requires one deletion, distance = 1)
    • Substitution – replacing a character with another (e.g., changing ship to shop requires one substitution, distance = 1)

    For example, transforming Sam into Adam requires two operations: inserting “A” at the beginning and replacing “S” with “d”, so the editing distance is 2.

    Levenshtein assigns equal weight to all operations. Variants such as Damerau-Levenshtein also account for transpositions of adjacent characters.

    Complete LivoLINK translation software

    Application of editing distance in translations

    Editing distance, although originating in maths and computer science, has practical applications in linguistics and translation. Modern translation tools use it extensively.

    Levenshtein distance in CAT tools

    CAT tools do more than match strings by editing distance; they use algorithms that consider context and subject matter. However, the foundation remains editing distance. For translators, this means receiving accurate suggestions and matches, enabling faster and more efficient translation.

    Editing distance and TM quality

    Tools using editing distance verify translation memory (TM) content, improving term and word matches. This reduces the risk of inconsistent terminology and allows translators to work more efficiently, especially with systems like LivoCAT.

    Editing distance and MT quality

    Machine translation (MT) produces quick translations but quality may vary. Editing distance metrics help translators evaluate MT output and estimate the effort needed to refine it.

    What is the significance of editing distance for translators?

    The metrics described above, developed decades ago, still help translators work on a wide variety of texts. Their core purpose is to compare two texts.

    One of these is a reference text (created “traditionally” by a human translator), while the other is a machine translation (produced by an algorithm or neural engine). Analysing both texts using editing distance helps estimate how much effort and time will be required to refine the translation to make it accurate.

    Editing distance metrics provide an overview of the translator’s workload by calculating the number of edits (operations) needed to transform one string of characters into another. These include:

    • character substitutions
    • character deletions
    • character insertions
    • character transpositions

    Comparing two texts in this way supports the post-editing process. Using editing distance to assess MT quality also affects the productivity of translators and proofreaders. It helps select the most suitable engine for a particular type of text and better manage project work, knowing how much time will be needed for completion.

    Editing distance in translations – summary

    Although machine translations (MT) are becoming more accurate and “natural,” human review, corrections and quality assessment remain essential.

    Editing distance metrics developed by linguists, such as Levenshtein, Hamming, or Damerau-Levenshtein, indicate how many operations are needed to turn one word (e.g., poorly translated) into another (e.g., correctly translated).

    Editing distance is expressed numerically, providing a clear, objective and readily accessible measure for translators and proofreaders. The higher the editing distance, the lower the quality of the translation and the more work the translator will face. This allows translation agencies or individual translators to plan project execution more efficiently and select the most suitable translation solutions, including machine translation engines. Translators can choose the tools that deliver the best results for specific types of texts.

    Why editing distance matters for translators

    Editing distance is highly valuable for translators who want to optimise their workflow. This measure of string differences is used in modern tools that assist translation. It enables automation and thus speeds up translation processes. Tools also help translators and project managers understand the amount of effort required to translate a text and complete a large project.

    The development of technologies based on editing distance does not replace human specialists. Algorithms perform tedious calculations and searches, while humans contribute creativity, cultural sensitivity and deep contextual understanding. Therefore, the translator of the future is a professional who not only knows languages but can also effectively leverage advanced linguistic solutions that rely, among other things, on Levenshtein editing distance.

    How editing distance supports modern translations

    Although rooted in mathematical concepts, editing distance has become the foundation of modern translation systems. It measures similarity between character strings and underpins CAT tool functions. Software using editing distance supports proofreaders and helps assess the quality of MT-generated texts. This measure allows translators to focus on what matters most – delivering high-quality translations – just as other software supporting translation workflows, such as CRM and TMS systems, does.

    POWIĄZANE WPISY: