Text edits

Whenever the system or the user modifies a source text or a translation in a segment, Wordbee keeps track of the previous text version. Every text change is kept in the history of a segment.

With the present API you can now have the system calculate the effort involved with all those edits done. Changing a typo requires less effort than amending a machine translation, which is probably also less effort than rewriting a complete sentence or typing a translation from scratch. The effort is measured by the (Levenshtein-) Edit Distance, see https://en.wikipedia.org/wiki/Levenshtein_distance

The API produces a report that shows the total effort by language, by user or other criteria. It further shows the effective source words/chars or translated words/chars that had been worked on. For example if there are 10 source words and a post-editor amended a small typo (suppose a 5% change of the text), the effective words is just 0.5 words in total.

Read more

See the API documentation:

Use cases

• Find out which user edited how many texts and also how much (small or big edit)

• Find out how many total edits had been done in a project in each language

• Get word counts adjusted by the amount of editing. Translation from scratch: Count words in full. Post editing: Adjust to amount of editing done.

• Use figures to calculate costs based on the exact effort that was put into editing: 1 character typed counts less than rewriting a bad translation.

• Get adjusted counts for both source words/chars as well as target words/chars

• Identify source text changes. This is specifically interesting with https://wordbee.atlassian.net/wiki/spaces/API2/pages/491596