6.1 KiB
Contributing guidelines for Chinese-to-English
We are working on translating "Hello Algo" from Chinese to English with the following approach:
- AI translation: Carry out an initial pass of translations using large language models.
- Human optimization: Manually refine the machine-generated outputs to ensure authenticity and accuracy.
- Pull request review: The optimized translation will be double checked by the reviewers through GitHub pull request workflow.
- Repeat steps
2.
and3.
for further improvements.
Join us
We're seeking contributors who meet the following criteria.
- Technical background: Strong foundation in computer science, particularly in data structures and algorithms.
- Language skills: Native proficiency in Chinese with professional-level English, or native English.
- Available time: Dedicated to contributing to open-source projects with a willingness to engage in long-term translation efforts.
That is, our contributors are computer scientists, engineers, and students from different linguistic backgrounds, and their objectives have different focal points:
- Native Chinese with professional working English: Ensuring translation accuracy and consistency between CN and EN versions.
- Native English: Enhance the authenticity and fluency of the English content to flow naturally and to be engaging.
Don't hesitate to join us via WeChat krahets-jyd
or on Discord!
Translation process
Important
Before diving in, ensure you're comfortable with the GitHub pull request workflow and have read the "Translation standards" and "Pseudo-code for translation" below.
- Self-assignment: Visit GitHub projects to select an unclaimed task and mark it as
In Progress
. - Translation: We encourage preserving the original meaning while ensuring the translation is natural and fluent.
- Peer review: Please carefully check your changes before submitting a Pull Request (PR). After approval by two reviewers, it will be merged into the project.
Translation standards
Tip
The "Accuracy" and "Authenticity" are primarily handled by native Chinese speakers and native English speakers, respectively.
In some instances, "Accuracy (consistency)" and "Authenticity" represent a trade-off, where optimizing one aspect could significantly affect the other. In such cases, please leave a comment in the pull request for discussion.
Accuracy:
- Maintain consistency in terminology across translations by referring to the Terminology section.
- Prioritize technical accuracy and maintain the tone and style of the Chinese version.
- Always take into account the content and context of the Chinese version to ensure modifications are accurate and comprehensive.
Authenticity:
- Translations should flow naturally and fluently, adhering to English expression conventions.
- Always consider the context of the content to harmonize the article.
- Be aware of cultural differences between Chinese and English. For instance, Chinese "pinyin" does not exist in English.
- If the optimized sentence could alter the original meaning, please add a comment for discussion.
Formatting:
- Figures and tables will be automatically numbered during deployment, so DO NOT manually number them.
- Each PR should cover at least one complete document to ensure manageable review sizes, except for bug fixes.
Review:
- During the review, prioritize evaluating the changes, consulting the surrounding context as needed.
- Learning from each other's perspectives can lead to better translations and more cohesive results.
Translation pseudo-code
The following pseudo-code models the steps in a typical translation process.
def optimize_translation(markdown_texts, lang_skill):
"""Optimize the translation"""
for sentence in markdown_texts:
"""Accuracy is handled primarily by native Chinese speakers"""
if lang_skill is "Native Chinese + Professional working English":
if is_accurate_Chinese_to_English(sentence):
continue
# Optimize the accuracy
result = refine_accuracy(sentence)
"""
Authenticity is handled primarily by native English speakers
and secondarily by native Chinese speakers
"""
if is_authentic_English(sentence):
continue
# Optimize the authenticity
result = refine_authenticity(sentence)
# Add comments in the PR if it may break consistency
if break_consistency(result):
add_comment(description)
pull_request = submit_pull_request(markdown_texts)
# The PR will be merged after approved by >= 2 reviewers
while count_approvals(pull_request) < 2:
continue
merge(pull_request)
The following pseudo-code is for the reviewers:
def review_pull_requests(pull_request, lang_skill):
"""Review the PR"""
# Loop through all the changes in the PR
while is_anything_left_to_review(pull_request):
change = get_next_change(pull_request)
"""Accuracy is handled primarily by native Chinese speakers"""
if lang_skill is "Native Chinese + Professional working English":
# Check the accuracy(consistency) between CN and EN versions
if is_accurate_Chinese_to_English(change):
continue
# Optimize the accuracy(consistency)
result = refine_accuracy(change)
# Add comments in the PR
add_comment(result)
"""
Authenticity is handled primarily by native English speakers
and secondarily by native Chinese speakers
"""
if is_authentic_English(change):
continue
# Optimize the authenticity if it is not authentic English
result = refine_authenticity(change)
# Add comments in the PR
add_comment(result)
approve(pull_request)