Copy of LQA

Best Practices
1. All errors noted in LQA must be accompanied by a comment. Do not leave the comment cell blank.
2. Your LQA feedback will be reviewed by non-native speakers as well as translators, so it is important that you write
your comments with this in mind. Write in clear, professional English. A non-native speaker must be able to
understand the essence of the error without having to refer to the original translation and your suggested
correction.
3. Avoid emotional language in your comments. Comments should be as objective as possible.
4. Comments should not be overly lengthy. 1-2 sentences is sufficient.
5. You do not need to reference the error category in the comment. For example, you do not need to write
“Grammar error: typo in the word ‘shop’”. “Typo in the word ‘shop’” would be sufficient here.
6. When reporting a major or critical error, we need to understand the impact of the error on meaning. Please
explain in your comment how the error in the translation has changed the meaning of the source. A back-
translation can sometimes be useful in these cases.
7. Errors relating to style, particularly unnatural translations, can be difficult to explain. In these cases, try to
consider the cause of the unnatural translation. Has it been translated literally? Is the word order following the
English structure too closely? Has the translator used some uncommon or archaic terms?
Common Error Miscategorization

1. The “Improvement” level also carries some weight and therefore impacts the final score. For linguistic issues LQA
Reviewers do not want to penalize, please leave the severity blank in the scorecard.
2. Untranslated segments:
- For UI LQA: these should be given no error or severity in the scorecard
- For Leo Tests and Senior Linguist Probation Tests: these should be marked as “Accuracy” errors. The severity will
depend on its impact on the text - if there is a loss of meaning due to the untranslated segment, it should be major
(or critical if the loss of meaning has safety, legal or financial implications); if not, it should be minor.
3. Terminology’ refers to translations that do not follow a) the glossary in Admin Panel, or b) standard industry
terminology. This category should not be used to mean ‘errors on the word level’ e.g. mistranslation of a word,
incorrect spelling of a word.
Examples
Good Comment Bad Comment
(please avoid writing comments like these)
Translation has omitted ‘but not limited to’. Omission.
The translator has translated the source word Unnatural translation.

for word, so it is following the English structure
too closely, causing the translation to be
unnatural.
‘Face recognition’ is translated as ‘Facebook Target includes a major mistranslation.

recognition’.
‘Action’ has been translated in the wrong कार्य
context. Here, it means कारबाही.
Incorrect use of pronouns. Grammatical error.
Please use the more inclusive term for

"housecleaner", femme (or homme) de I'm shocked that the translator used such an
ménage. "La bonne" is closer to "maid", which old-fashioned term here. Nobody uses "la
has a negative connotation. bonne" for housecleaner anymore.
The keigo (honorific) suffix 様 (sama) is too

formal for training manuals. Use さん (san), as
stated in the style guide (voice/tone). The keigo suffix 様 is too formal.
Error Categories Description
Used to mark inappropriate language:

- Hate speech
- Violent language
Abusive
- Explicit language
- Language that appears to have a political motivation or connotation
Used to mark meaning issues, including:

- Mistranslations — when the wrong message is conveyed in the translation
- Omissions — when essential elements in the source text are missing in the
Accuracy translation
- Additions — when unnecessary elements are present in the translation,
which are not originally present in the source
This category is applied when edits are needed because of source content
issues. These edits do not count as errors and have no severity. This can be
caused by:
- Incomprehensible or ambiguous content
- Inaccurate product information, e.g. outdated or inconsistent product names,
Source Error
discrepancy with the product UI, products that are not offered in a specific
market
- Segmentation issues
Used to mark the following format issues:

- Missing, misplaced or wrong formatting tags, placeholders (tokens) or
variables
- Omitted structural, non-textual elements, e.g. bullet points
Format
- String-length violations
- Links not working (for testing tasks)
- Truncations and overlaps (for testing tasks)
Used to mark grammatical errors:
- Grammar, spelling and punctuation errors
Grammar - Wrong syntax that violates the target language rules
- Violations of grammar, spelling and punctuation guidelines in the Facebook
Style Guide
Used to mark the following stylistic issues:

- Noncompliance with tone and style guidelines provided in the Facebook Style
Guide, reference materials or project instructions
- Low readability resulting from word-for-word translation, awkward syntax,
Style
verbosity, unnecessary repetition or other fluency issues impacting the user
experience
- Inappropriate tone or style for the given context, e.g. inappropriate level of
formality, colloquial, or archaic words or expressions
Used to mark the following terminology issues:

- Noncompliance with the Facebook glossary or other applicable terminology
reference materials
- Noncompliance with standard industry terminology
Terminology
- Inconsistent use of the same terms throughout a document or project
- Noncompliance with publicly accessible third-party product terminology or UI
elements (ex., MS Windows and Android OS)
Preset severity level: None. This category is applied when the UI content was
fully untranslated. This category should be used for UI localization testing only,
and should not be used for non-UI translation projects. Accuracy error
Untranslated category should be applied when the content was partially or fully
untranslated where it should have been translated. Terminology error
category should be applied when a term or a product name was left
untranslated where it should have been translated.
Why we don't have 'Consistency'?

Please choose a specific error category instead. If an error is about a term consistency, choose
'Terminology' instead. If it's about a stylistic consistency issue, choose 'Style'.
Error Severities Description
Errors that may carry safety, legal or financial implications,

damage the company’s reputation, negatively
modify/misrepresent the functionality of a product or service,
or which could be seen as offensive.
Example:
Critical
Source: By clicking Agree, you Agree to our Terms of Use and
that you have read our Privacy Policy, including our cookie
use.
Target: By clicking Agree, you Agree to our Terms of Use and
that you have read our Privacy Policy, excluding our cookie
use.
Errors that may confuse or mislead the user or hinder proper

use of the product/service due to significant change in
meaning, or because errors appear in a visible or important
part of the content.
Major
Example:
Source: More than 60% of stories videos are played with
sound on.
Target: More than 60% of stories videos are played with
sound off.
Errors that do not lead to loss of meaning and would not

confuse or mislead the user but would be noticed, would
decrease stylistic quality, fluency or clarity, or would make
the content less appealing.
Minor
Example:
Source: The number of times your video has been watched
for at least 10 seconds.
Target: The number of times your video has been watched for
at least 10 second.
Edits that show the current translation is acceptable but not

optimal. Improvements do count as very slight errors (penalty
is negligible) because this type of edit improves the
translation quality.
Improvement
Example:
Source: Mobile moves fast.
Target: Mobile moves at rapid speed.
Applies to instances of the same error on additional segments
Repeated
within the same task. See instructions on how to use below.
Quality Scorecard
Client Scorecard Name Start Date Content Type

11/1/2024 Standard
Source Language Target Language Wordcount Quality Model Content Category
en_US tr_TR 573 General QM #N/A (Not Defined)
Final Errors Topology

Category Critical Major Minor Improvement
Abusive 0 0 0 0
Accuracy 0 0 0 0
Source Error 0 0 0 0
Format 0 0 0 0
Grammar 0 0 0 0
Style 0 0 0 0
Terminology 0 0 0 0
Untranslated 0 0 0 0
Total 0 0 0 0
Linguistic Evaluation
Reviewer Initial Score Final Score
#REF! #REF!
Vendor Type Initial Rating Final Rating
#REF! #REF!
N/A Repeated Total

0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
Source Target Edited Target
Revision
Overall Evaluation
Initial Category
Revision
Overall Evaluation
Initial Severity Comment

Copy of LQA

Uploaded by

Copyright:

Available Formats

Copy of LQA

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Copy of LQA

Uploaded by

Copyright:

Available Formats

Best Practices

3. Avoid emotional language in your comments. Comments should be as objective as possible.

4. Comments should not be overly lengthy. 1-2 sentences is sufficient.

Common Error Miscategorization

The translator has translated the source word Unnatural translation.

‘Face recognition’ is translated as ‘Facebook Target includes a major mistranslation.

Please use the more inclusive term for

The keigo (honorific) suffix 様 (sama) is too

Used to mark inappropriate language:

Used to mark meaning issues, including:

Used to mark the following format issues:

Used to mark the following stylistic issues:

Used to mark the following terminology issues:

Why we don't have 'Consistency'?

Errors that may carry safety, legal or financial implications,

Errors that may confuse or mislead the user or hinder proper

Errors that do not lead to loss of meaning and would not

Edits that show the current translation is acceptable but not

Client Scorecard Name Start Date Content Type

Final Errors Topology

N/A Repeated Total

Initial Severity Comment

You might also like