Default word count settings

When you access the settings for the first time, you will see the following fuzzy match intervals set by default:

You can adjust the default configuration to match the discounts on the fuzzy matches you negotiated with each of your suppliers. Here is an example:

• Interval 1: 100% - 110%

• Interval 2: 95% - 99%

• Interval 3: 85% - 94%

• Interval 4: 75% - 84%

• Interval 5: 50% - 74%

The system will show the number of source words that were found in a translation memory with a certain degree of similarity. The different percentages or intervals indicate the amount of editing required. For example, matches between 85% and 110% are considered high fuzzy matches as little editing is required. Therefore, translators could apply various discounts on their full rate per word. Matches lower than 75% require more editing effort and even complete rewriting. In this case, no discounts apply.

The match values need to correspond to the discounts on the fuzzy matches you negotiated with your suppliers. To check the pricelists, go to the Suppliers dashboard> select the supplier you would like to use and then open the Prices tab.

Note that Wordbee allows you to define discounts on both fuzzy matches and pre-translation matches. In the above example, the same discounts have been applied for both match categories.

100% - 110% matches

110% stands for perfect or in-context matches: not only the source segment you are translating is similar to the one stored in the TM but also the context before and after the segment. In principle, perfect matches do not require any editing or review by linguists.

100% matches are exact matches: the active segment in the Editor is exactly the same as the one stored in the TM.

In contrast with the perfect matches, 100% matches might need slight review.

95% - 99% fuzzy matches

There are slight differences between the source segment you are currently translating and the one stored in the TM, such as capitalization, number, formatting, spacing.

85% - 94% fuzzy matches

The difference might consist of one full word between the source segment you are translating and the one stored in the TM.

75% - 84% fuzzy matches

There are more differences in words and formatting.

50% - 74% fuzzy matches

Matches below 75% require more editing or translation from scratch.

Pre-translation from memories

If you select this option, the system will pre-translate segments from all project's resources, namely the project memory, translation memories or terminology databases.

The threshold specifies whether exact or fuzzy matches are used to pre-translate the segments.

Pre-translation from project memory

If you select this option, the system will pre-translate segments only using the project memory. Should there be other translation memories attached to the project, the matches from the project memory will be proposed first.

If this option is not selected, the project memory will be used to pre-translate repetitions only.

Pre-translation by machine translation (MT)

If you select this option, the system will pre-translate the source document (s) with the help of the Machine Translation when there are no matches coming from the Translation Memories.

Count already translated segments as pre-translations

If you select this option, a word count considers already translated segments as pre-translated (e.g. if a document is partially translated and a word count is re-run). This also applies to initial counting of multi-lingual Excel sheets, POT or XLIFF files which often already contain translations.

Lock pre-translations (make read-only)

This option allows you to specify if pre-translations are automatically made read-only in the translation editor - indicated by the 'locked' icon.

Usually, project managers lock the perfect matches as they do not require any editing work. If the segments are locked, the linguists are not able to modify them.

Count pre-translations as “not to translate” (do not cost)

Use this option to count all or specific pre-translations as "not to translate". Any such counts are ignored by cost calculation. In addition, you might want to make these texts "read only" in the editor with the option above.

Substitute dates and numbers

If a very similar memory hit is used for pre-translation, dates and numbers are replaced with the corresponding dates and numbers from the source text (where these do differ).

Fix markup during pre-translation

If set, the formatting contained in a memory hit used for pre-translation is automatically adjusted: Inline tags (markup) are added, renamed or removed so as to match the translated text's formatting. Recommended setting.

Enable repetition counting

If you enable this option, the system will identify the repetitions within a project's documents and counts these as "fuzzy matches or repetitions". Repetitions are identical or similar segments that occur again and again within one document or across all documents in the project.

Identical or similar repetitions

The system either counts identical repetitions only (100% to 110%) or extends the search to similar matches (within the fuzzy intervals indicated above). If you use this option, the system will count only the identical repetitions.

Translated or untranslated repetitions

When counting a document, the system counts translated and untranslated repetitions not only within the document itself but also across all documents in the project. If you use this option, the system will count only those repetitions that have been translated.

Count alphanumeric sequences as words

If ticked, then strings like "A13" or "F57017" are considered.

Count any numbers as words

If ticked, numbers are counted as word. Untick to not count numbers.

Count well delimited numbers as words

If set, numbers will be systematically counted as words even if these are not separated by spaces, commas or other word separators. For example: "Line13" will count 2 words.

Count any punctuation characters as words

If ticked, any punctuation character is counted as a word whether such characters appear within a word or not.

Count any symbols as words

If ticked, any symbols that are well delimited by word separators are counted as words. For example, if not ticked then "alpha *" returns 2 words but "alpha*" returns just one. If ticked, both examples would return 2 words.

Count well delimited symbols as words

If ticked, then symbols that are delimited by word separators (spaces, full stops, etc.) are counted as words.

Count dates as one word

If ticked, then dates are counted as one single word.

Count measurements as one word

If ticked, then measurements (1 cm, 100 kg, etc.) are counted as one word.

Ignore inline tags

If ticked, then formatting tags are not counted. Otherwise, such tags will be counted as words.

Count Asian characters as words

If ticked, each Asian character will be counted as an individual word.

Adjust to standards

If ticked, counting is optimized to be as compatible as possible to market leading word count standards.