Description of checks in the Quality Assurance (QA) profile
- 1 Introduction
- 2 Description of QA checks
- 2.1 Numbers
- 2.2 Punctuation
- 2.2.1 “Missing space after punctuation marks” QA check
- 2.2.2 “Multiple punctuation marks in a sentence” QA check
- 2.2.3 “Multiple spaces in a sentence” QA check
- 2.2.4 “Spaces before punctuation marks” QA check
- 2.2.5 “No spaces before punctuation marks required for French (Switzerland)” QA check
- 2.2.6 “Spaces around inlines” QA check
- 2.2.7 “Spaces around double byte characters” QA check
- 2.2.8 “Trailing tabs or spaces at the beginning or end of a sentence” QA check
- 2.2.9 “Uneven number of opening and closing brackets” QA check
- 2.3 Language
- 2.4 Terminology
- 2.5 Matching
- 2.6 Other
- 2.7 Regular expressions
Introduction
This article provides a detailed description of all checks that are available in the Quality Assurance (QA) profile, in the XTM Cloud UI.
Before reading the article below, ensure that you have a good understanding of how to set up and use the QA feature, which is described in the following articles: |
|---|
Description of QA checks
In the QA profile, the checks are categorized by purpose. Consequently, this section has a similar structure.
IMPORTANT!
Keep in mind that the scope of QA checks depends on the connected spell checker and its availability for a particular language.
Numbers
“Missing or incorrectly formatted number” QA check
Checks that a number displayed in a source segment has been localized in compliance with the grammar rules of a particular target language. If the check detects an inconsistency, it suggests the correct form.
“Numbers inside words” QA check
Checks if there are any unexpected numbers inside words in a target segment. Each affected word is highlighted.
“Extra number in target text” QA check
Checks if a target segment contains more numbers than its source segment. Each additional target-segment number is highlighted.
Punctuation
“Missing space after punctuation marks” QA check
Checks if there are any missing spaces after punctuation marks such as a comma or full stop, in the target segment. Each occurrence is highlighted.
“Multiple punctuation marks in a sentence” QA check
Checks if there are any repeated punctuation marks next to one another, in the target segment. Each occurrence is highlighted.
“Multiple spaces in a sentence” QA check
Checks if there are any repeated spaces (blanks) in the target segment. Each occurrence is highlighted.
“Spaces before punctuation marks” QA check
Checks if there are any redundant spaces before punctuation marks in the target segment. Each occurrence is highlighted.
“No spaces before punctuation marks required for French (Switzerland)” QA check
In standard French, there MUST BE blank spaces before certain punctuation marks: colon, semicolon, exclamation mark or question mark.
In French (Switzerland), there MUST BE NO blank spaces before these particular punctuation marks.
This QA check verifies that, in a target text written in French (Switzerland), there are no blank spaces before the punctuation marks listed above. Each occurrence is highlighted.
“Spaces around inlines” QA check
Checks if there are any redundant spaces around inline tags in the target segment, compared to the source segment.
“Spaces around double byte characters” QA check
Checks if there are any redundant spaces between double-byte characters in the target segment.
Each occurrence is highlighted.
A double-byte character set (DBCS) is a character encoding in which each character, or the large majority of them, is represented by two bytes instead of one. This is necessary for languages like Chinese, Japanese or Korean that have thousands of characters and symbols that cannot be represented using a single-byte system such as ASCII.
“Trailing tabs or spaces at the beginning or end of a sentence” QA check
Checks if there are any redundant spaces at the beginning or end of a target segment.
A trailing space is all whitespace located at the end of a line, without any other characters following it. This includes spaces (blanks), tabs \t, carriage returns \r, etc.
“Uneven number of opening and closing brackets” QA check
Checks if there is any difference in the number of opening and closing brackets contained in a source segment and its target segment.
Language
“Check non-translatable text in curly brackets” QA check
Check whether any text in curly brackets, that should not be translated, in a target segment, has actually been translated.
“Forbidden characters” QA check
Checks whether any forbidden characters have been inserted in a target segment. To set specific characters as forbidden, select the Forbidden characters checkbox and then enter the characters in the text field that is displayed beneath the checkbox.
“Identical source and target” QA check
Checks whether the texts in the source and target segments are the same.
“Repeated words” QA check
Checks if there are any repeated words in a target segment. Each occurrence is highlighted.
“Spelling errors” QA check
Checks for any spelling mistakes in a target segment. Each misspelt word is highlighted.
“Check grammar and style” QA check
This QA check verifies the translation in terms of its overall grammatical and stylistic quality, and detects potential errors. Each error is highlighted.
To learn more about this particular type of QA check, read the following article: Quality Assurance (QA) grammar checker.
“Language Guard” QA check
Checks whether any offensive or non-inclusive language has been used in a target segment. This AI-driven functionality helps you uphold a professional and inclusive brand image that resonates positively with your target audience. Once an issue of this type is detected, you are given suggestions on how to restructure the sentence to help address the issue. Of course, accuracy might vary, so Linguists must always review the flagged words in context.
IMPORTANT!
This QA check is not available out of the box! Your XTM Cloud instance must have an active OpenAI GPT subscription in place. For details, see the following article: OpenAI GPT MT integration.
This QA check is supported for all language variants that are supported by OpenAI GPT.
Terminology
“Forbidden terms used” QA check
Checks if a forbidden translation of a particular term has been used in a target segment. This functionality takes into account translations, using XTM Terminology, that have been marked as Forbidden in the XTM Cloud UI.
“Rejected terms used” QA check
Checks if a rejected translation of a particular term has been used in a target segment. This functionality takes into account translations, using XTM Terminology, that have been marked as Rejected in the XTM Cloud UI.
“Translation of a term not found” QA check
Checks if a translation is available for a particular term, in the XTM Terminology, but that translation is missing in the target segment.
“Translation of term does not exist in terminology” QA check
This QA check verifies if the translation of a particular term into the project’s target language exists in XTM Terminology. If it is not, an information message is displayed in the QA docked panel.
Matching
“Unedited fuzzy TM match” QA check
Checks whether a Fuzzy Match has been applied in the target segment, without being edited. If this is the case, an information message is displayed in the QA docked panel. Of course, an unedited fuzzy TM match need not always necessarily be an incorrect translation, but it still needs to be checked by a Linguist.
“Unedited fuzzy repetition” QA check
Checks whether a Fuzzy Repetition has been applied in the target segment, without being edited. If this is the case, an information message is displayed in the QA docked panel. Of course, an unedited fuzzy repetition match need not always necessarily be an incorrect translation, but it still needs to be checked by a Linguist.
Other
“Order of inlines is changed or content is missing between inlines in target segment” QA check
Checks that all the inline tags that exist in the source segment are in the correct sequence, in the target segment, and whether there is any missing text between those inline tags. Even if this check generates an issue, there might be no inline tag errors: the placement of inline tags also depends on word-order differences between source and target languages, for example.
Regular expressions
Regular expressions can also be used to perform an in-depth quality assurance check on target segments in XTM Workbench. To learn more about this functionality in detail, read the following article: Regular expressions (regex) in a QA profile.