You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This document is intended to help you quickly understand the PCL i18n Project workflow and point out common translation issues to contribute more efficiently and accurately. Whether you're a newcomer to translation projects or an experienced translator, this guide provides clear operational steps and practical advice. By following the processes outlined here, you can ensure the quality and consistency of your translations while avoiding common errors and misunderstandings. We hope this guide becomes a valuable assistant in your contributions to the PCL i18n Project!
Overall Translation Workflow
To ensure the quality and efficiency of the translation process, we have established the following workflow that balances accuracy, adaptability, consistency, and conciseness. Considering that directly translating from Chinese into certain languages (such as Western European languages) may result in lengthy or poorly structured text, we prioritize translating into English first, refining the expression, and then using it as the base version for other languages.
Therefore, it is essential to understand and adhere to the established workflow during the translation process to avoid non-standard operations. For example: submitting translations in English (United Kingdom) that are entirely different from English (US) is not acceptable; derived language translations should not be created directly from Chinese; and relying solely on basic simplified-traditional conversion tools for language transformation is not permitted.
The workflow is as follows:
1. Upstream Repository Update & Language File Preparation
When the main project repository (tangge233/PCL2/tree/Language) is updated, we will perform the following actions:
Synchronize the updated content.
Resolve any conflicts.
Add new default Chinese strings to all language files.
2. First-Level Translation: Chinese → Primary Languages
Newly added or updated Chinese strings are first pushed to the Weblate platform for the initial round of translation. The target primary languages include:
English (US)
日本語 (日本)
한국어 (대한민국)
文言 (華夏)
繁體中文 (台灣) - converted using 繁化姬 (ZhConvert) and manually proofread
梗体中文 (天朝)
焱暒妏 - processed using an automatic converter
3. Second-Level Translation: Primary Languages → Derived Languages
Once the first-level translations are complete, their content will serve as the base for translating into derived languages. The corresponding relationships are as follows:
Based on English (US):
Ελληνικά (Ελλάδα)
English (United Kingdom) - with adjustments to vocabulary and punctuation style
Español (España)
Français (France)
Русский (Россия)
Slovenčina (Slovensko)
Based on 繁體中文 (台灣):
繁體中文 (香港特別行政區) - converted using 繁化姬 (ZhConvert) and manually proofread
4. Translation Synchronization & Submission
Once all first-level and second-level translations are complete, the changes made on Weblate will be committed to the repository.
To aid in understanding, a clear flowchart is provided:
The previously introduced workflow focused on translating entire languages. This section provides a detailed explanation of how to translate and handle individual strings step by step, aiming to improve both the efficiency and quality of translations.
1. Submitting Translation Suggestions
This process applies to strings that have not yet been translated or approved.
Identify the Source Language: Determine the language the string originates from, and refer to the “Translation Workflow” section to find its corresponding parent language.
Initial Translation: Use AI tools (ChatGPT series recommended) or translation software to create a first draft. Ensure the translation is fluent and semantically accurate while maintaining efficiency.
Manual Polishing: Refine the draft translation to ensure consistency with the original text and avoid over-interpretation or misrepresentation.
Length Evaluation & Optimization: Locate the string in the PCL interface and assess whether the translated content fits well within the UI constraints. If needed, simplify expressions and remove redundant words.
Submit Suggestion: Once the translation is accurate and appropriately formatted, submit it as a suggestion.
After saving the comment, wait for admin review. To expedite the process, it is recommended to initiate a discussion in the QQ group or Discord channel.
3. Reviewing Translation Suggestions
This applies to strings that have been translated but not yet approved.
Review Suggested Translations: Carefully examine each suggestion to ensure the meaning is clear, the grammar is correct, and the translation fits the context.
Use Proofing Tools: Utilize AI tools to assist with grammar checks and to ensure the translation is aligned with natural language usage.
Length Evaluation: Check the translation within the PCL interface to confirm it fits appropriately. If it's too long, simplify the wording as needed.
Approve Translation: Set the Review Status to Approved and click save.
graph LR
C1[Review Suggestions] --> C2{Grammar Check}
C2 -- Grammar Issues --> C5[Accept & Edit]
C2 -- No Issues --> C3{Length Evaluation}
C3 -- Too Long --> C5[Accept & Edit]
C3 -- Appropriate --> C4[Set as 'Approved' and Save]
Loading
4. Re-reviewing Approved Translations
This process applies to translations that have already been approved but are found to contain errors or have potential for improvement.
Read Comments: Review comments submitted by other contributors to understand the reported issue. If anything is unclear, you may ask follow-up questions directly under the original comment.
Evaluate Validity: Assess whether the suggestion is reasonable. If not, reject the suggestion, provide an explanation, and mark the comment as Resolved.
If Edits Are Needed: Change the Review Status to Needs Editing and save it. The string will then enter a pending modification state.
Reach a Consensus: Discuss the issue in the comments section or through community channels (e.g. group chat) to reach a consensus. If agreement is achieved, revise the translation and proceed through the “Reviewing Translation Suggestions” process again. If no consensus is reached, reject the suggestion and mark the comment as Resolved.
graph LR
D1[Read Comments &
Understand Suggestion] --> D2{Reasonable?}
D2 -- Yes --> D3[Mark as “Needs Editing”
and Save]
D3 --> D4{Consensus Reached?}
D4 -- No --> D6[Reject Suggestion &
Mark as “Resolved”]
D4 -- Yes --> D5[Re-translate &
Run Review Process]
D2 -- No --> D6
Loading
Special Language Translations
Some languages require slightly different translation procedures compared to others. Below are specific guidelines for handling special languages:
English (United Kingdom)
When converting American English to British English, standard translation tools or AI services may not accurately handle differences in spelling and vocabulary. Therefore, we recommend using professional spelling checkers or converters to proofread the translations and ensure they conform to British English conventions.
Microsoft Word: Built-in spelling and grammar check functionality
繁體中文 (台灣)
Standard translators, AI tools, or simplified-traditional converters typically only perform basic character shape conversions and often fail to accurately address vocabulary and semantic differences. For example:
网络 → 網路
设置 → 設定
内存 → 記憶體
Additionally, due to the simplification process where multiple traditional characters were merged into a single simplified form, it's often difficult to reverse-convert them correctly, resulting in ambiguity. Examples include:
髮, 發 → 发
復, 複, 覆 → 复
臺, 檯, 颱 → 台
If simple find-and-replace tools are used on top of this, it can lead to laughable misinterpretations (such as “海記憶體知己”).
To address these issues, we have selected ZhConvert (繁化姬) as our standard conversion tool. It not only converts Simplified Chinese to Traditional Chinese with high accuracy but also handles most of the differences in vocabulary and usage, meeting practical translation needs to the greatest extent possible.
After conversion, a quick manual review and correction of a few terms will suffice.
How to Use
ZhConvert (繁化姬) is extremely easy to use. Simply paste the Simplified Chinese text into the input box and click the convert button to quickly generate the corresponding Traditional Chinese output. We recommend following these steps:
Enable the following recommended modules to improve conversion accuracy and coverage:
修飾句子 (Sentence Enhancement)
單位轉換 (Unit Conversion)
專有名詞 (Proper Nouns)
引號 (Quotation Marks)
正、異體字 (Standard & Variant Characters)
重複字修正 (Repeated Characters Correction)
重複字通用修正 (General Repeated Characters Fix)
電腦詞彙 (IT Terminology)
If needed, you may adjust the enabled modules based on the specific context, but it is strongly recommended to keep the above modules active for optimal quality.
Return to the 轉換 (Convert) tab.
Paste the Simplified Chinese text into the 轉換前文字 (Text Before Conversion) input box.
Click the 台灣化 (Taiwanize) button.
Copy the converted Traditional Chinese text from the 轉換後結果 (Converted Result) section.
繁體中文 (香港特別行政區)
Due to the limited support for 繁體中文 (香港特別行政區) in most mainstream translation tools, AI models also often struggle with accurately handling Hong Kong-specific expressions. Therefore, we similarly recommend using ZhConvert (繁化姬) for conversions.
The conversion process is nearly identical to that for 繁體中文 (台灣), with just two key differences to note:
The source language should be set to 繁體中文 (台灣)
In Step 6, click the 香港化 (Hong-Kongize) button instead of 台灣化 (Taiwanize).
⚠️ Important Notes:
Since ZhConvert does not fully cover all nuances of Hong Kong-specific terminology, it is recommended to manually proofread and replace individual words after conversion to ensure that the final translation better reflects local usage. Additionally, be sure to always enable the 正、異體字 (Standard & Variant Characters) module, so the output conforms to the traditional character forms used in Hong Kong writing conventions.
焱暒妏
The translation of 焱暒妏 (Martian Script) requires a specialized converter. It is not an independent language in the traditional sense, but rather a stylized form of writing achieved by altering character shapes to create a unique visual effect.
We recommend using the following tool for Martian script conversion: Martian Text Generator. Simply paste the original text into the input box, and the tool will instantly generate the corresponding Martian-style text.
文言 (華夏)
WIP……
梗体中文 (天朝)
Based on the principle of quantum semantic superposition and the mechanism of non-linear symbolic dynamics, Meme-style Chinese reconstructs the metastable structure of traditional language matrices atop the topological foundation of classical Chinese characters. Specifically, this linguistic variant, under the constraint of the Fourier Semantic Conservation Law, evolves through the synergistic interaction of high-frequency meme resonance factors and discretized meme-cellular automata. Within a nonequilibrium pragmatics framework, it gives rise to a novel semantic superstructure characterized by fractal chaos. The core process involves a second-order harmonic oscillation of morphemes over the original Chinese base, injecting self-similar subcultural symbols through the entropy-expanding channel of internet culture.
As the name suggests, 梗体中文 (天朝) (Meme-style Chinese) is based on standard Chinese but infused with a rich mix of internet slang and meme elements. During the actual translation process, feel free to unleash your imagination and creativity—blend humor with abstraction, and don’t hold back from embracing the ✨chaotic energy✨ of meme culture. Let your inner gremlin shine through, yasss!!!
There are very few restrictions on what can be done in Meme-style translations—as long as the result isn’t excessively inappropriate, you’re encouraged to have fun and be as playful and expressive as you like.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Warning
内容为草案,随时可能修改。因此较早讨论内容可能与当前呈现内容有所出入,请仔细核对时间轴信息。
PCL i18n & l10n 项目指南
Milestone I18n
摘要
本提案旨在为 PCL-CE 的国际化(Internationalization, i18n)与本地化(Localization, l10n)策略确立具体实施路径。提案涵盖语言资源管理机制、键名规范制定、时间与数字格式适配、以及区域限制功能的设计思路。通过标准化语言资源与格式处理流程,本提案预期提升系统的多语言支持能力与全球用户的使用体验,并为未来多地区版本部署打下基础。
目标
本提案承接此前在 PCL 项目中通过 PR#4145 实施 i18n 的初步尝试,进一步在 PCL-CE 上推进 i18n 支持的系统化建设。基于过往工作经验与实践反馈,提案旨在明确 PCL-CE 项目中 i18n 与 l10n 工作的标准化实施范围与总体目标,确保多语言支持具备一致性、可扩展性与长期可维护性。
描述
本提案将围绕 PCL-CE 项目中 i18n 与本地化 l10n 的具体实施展开,涵盖语言资源的存储形式与存放位置、键值命名规范、区域化格式支持、区域判断方式,以及当前翻译语言的处理策略。
总体架构与资源结构
考虑到
.resx文件在进行批量操作时存在实际困难,项目当前计划继续采用.xaml文件来存储语言资源。同时,XAML 文件与 WPF 框架具有良好的兼容性,适合现有界面的调用方式,便于界面端直接引用与维护。目前,PCL-CE 的项目结构中包含以 C# 编写的 Core 代码库,以及以 VB.NET 编写的程序主体库。考虑到后续与语言相关的逻辑将更多集中在 Core 层,因此,语言资源文件及其配套的实现逻辑将主要在 Core 中开发与维护。
键值的命名规范
在早期 PCL Language 实践中,虽然语言键值按照模块划分,但未使用清晰的分隔符,导致部分键名语义模糊、难以理解,即便有注释也不易辨识。
本提案引入明确的键名分组结构,统一使用
.(dot)作为分隔符,建议格式为:该命名规范在保持模块化前提下,提高了可读性与可维护性,也有助于日后的审校与扩展。
需要注意的是,不同界面可能存在重复性文本,在组织过程中需谨慎评估重复内容的归属与去重策略。可参考 HMCL 项目的语言资源结构,其按照“功能”而非“位置”划分的方式,具有一定参考价值。
区域化展示
不同地区的用户在阅读时间、日期和数字时的偏好,或各地区不同机构所规定的标准表示方法各不相同。在早期的 PCL Language 实现中已涉及部分本地化处理,但未能充分利用 .NET 框架内建的功能。
因此,当前计划采用 .NET 框架提供的
CultureInfo实现时间与数字的本地化格式转换,以满足不同地区用户的阅读习惯,减少手动格式拼接所带来的错误与冗余。例如:
区域功能判断机制
为支持区域限制性功能,可引入地区识别逻辑。具体实现方式分为外部判定机制与本地环境参考。
外部判定机制借鉴了 SJMCL 项目的实践,通过访问
https://cloudflare.com/cdn-cgi/trace接口,获取连接状态数据,从中提取loc字段用于判断用户所处国家或地区。本地环境参考将会同步结合用户计算机系统的区域设置、语言设置等环境信息,作为辅助参考来源,以提供更可靠的区域识别手段。
示例获取逻辑如下:
翻译语言与引入方式
目前翻译计划中共包含 14 种语言,具体列表可参见主线的 PR#4145 内容。由于人力资源较为紧张,应优先复用原有高质量翻译资源,避免重复翻译与风格不一致。除此之外,对于人手不足的语种,可在人工校对前提下,适度借助大语言模型辅助生成翻译草稿。所有翻译内容应统一接入语言资源结构,确保整体一致性与后续可维护性。
除此之外,还应参考原有的 PCL i18n & l10n 项目指南
影响
在系统结构方面,语言资源的集中管理将促使 Core 层承担更多职责。原本分散在各界面模块中的语言文本将统一迁移至 Core 中维护,并由程序主体库进行调用。这一结构调整将影响部分模块的资源加载路径和引用方式,可能涉及调用逻辑和依赖结构的改动。此外,时间、日期和数字的显示将统一采用 .NET 提供的
CultureInfo格式化机制,原有使用手工拼接的展示代码将被替换为标准化调用方式。这一改变有助于提升一致性,但也要求对相关代码逻辑进行更新。在用户体验方面,多语言资源的统一加载和本地化格式的适配,将直接提升非中文语境用户的使用便利性和阅读体验。用户界面中的文本、时间、数字等内容将能够根据用户所处地区自动调整,体现更符合本地习惯的显示逻辑。
从开发与维护角度看,i18n 与 l10n 的全面引入将短期内提高开发复杂度和测试压力。开发者需要理解语言资源的结构、命名规范及其调用方式,在编码过程中额外关注语言键值的正确性与复用性。测试流程也需同步引入多语言覆盖场景,确保不同地区配置下的显示与行为一致。此外,项目需新增一套与语言资源相关的文档与协作流程,便于团队成员理解结构,也可规范翻译行为。
语言资源结构的标准化将提升社区翻译协作的门槛,贡献者需要理解键值命名逻辑与资源组织方式,这对初次参与者可能造成一定障碍。社区治理方面,项目或需要建立新的翻译审核流程和合并机制,以保证翻译质量和风格统一。
兼容性
本提案将在语言资源结构、格式化逻辑以及模块依赖上引入变动,可能影响现有代码的兼容性。原有分散的语言键值与硬编码字符串将被统一替换为集中管理的 XAML 语言资源,界面层需要调整对文本的引用方式,原有引用路径将失效。部分模块可能依赖旧有资源结构或手动拼接逻辑,在未完成同步迁移的情况下,会出现显示缺失或格式错误的问题。
日期和数字格式的显示将改为依赖 .NET 的
CultureInfo实现方式,这与之前使用固定格式或字符串拼接的处理方法存在行为差异。某些依赖具体格式字符串的功能逻辑(如导出文件命名、日志时间戳格式)在切换后可能产生不一致,需在接口调用处做兼容性适配。语言资源的命名规范将发生变化,原有不规范或无结构的键值将无法与新结构直接对应。对历史翻译内容进行复用时,如未同步重命名,可能导致内容映射错误或资源缺失。
替代方案
.resx
在语言资源的存储方式上,项目最初曾考虑采用 .NET 生态中较为标准的
.resx资源文件方案。该格式在 Visual Studio 中具备良好的集成支持,便于通过设计器进行可视化编辑,也方便与.Designer.cs自动生成类联动,具备较高的 IDE 兼容性。然而,
.resx文件在多语言资源批量操作方面存在较大限制,特别是在版本控制、批量比对与翻译协作场景下,难以高效维护。其结构基于 XML 且带有编码嵌套,人工编辑体验较差,不适合由非技术成员直接参与修改。同时,.resx与 WPF 的集成方式相对复杂,缺乏直接的数据绑定能力,需要额外封装中间层。基于上述考量,最终选择继续使用
.xaml作为语言资源的主要存储格式。该格式与 WPF 框架原生兼容,支持样式、模板与数据绑定场景下的语言切换,亦便于在 Core 层中集中维护资源结构。相较.resx,.xaml更适合现有架构和开发习惯,便于在不引入额外转换逻辑的前提下进行集成。附录
附录 1:PCL i18n & l10n 项目指南
PCL i18n & l10n 项目指南
Scroll down for the English version!!!
本文档旨在帮助您快速了解 PCL i18n 项目 的工作流程,并指出翻译过程中常见的问题,以便更高效、准确地参与贡献。无论您是初次接触翻译项目的新手,还是经验丰富的翻译人员,本指南都将为您提供清晰的操作步骤和实用的建议。通过遵循本文档中的流程,您可以确保翻译内容的质量与一致性,同时避免常见的错误和误解。希望本指南能成为您参与 PCL i18n 项目时的得力助手!
整体翻译工作流程
为确保翻译质量和流程效率,我们制定了如下翻译流程,兼顾准确性、适配性、一致性及精炼性。考虑到从中文直接翻译为某些语言 (如西欧语言) 可能会造成文本冗长、结构不佳等问题,我们优先将其翻译为英文,优化表达后再作为其他语言的基础版本。
因此在翻译过程中,务必了解并遵循既定的流程,以避免产生不规范的操作。例如:不应在 English (United Kingdom) 中提交与 English (US) 完全不一样的译名;不得直接使用中文译名进行派生语言翻译;也不可仅依赖基础的简繁转换工具进行语言转换。
流程如下:
1. 上游仓库更新与语言文件准备
当项目主仓库 (tangge233/PCL2/tree/Language) 更新后,我们将进行以下操作:
2. 第一级翻译:中文 → 主语言
新添加或更新的中文字符串首先被推送至 Weblate 平台,进行第一轮翻译。目标语言包括:
3. 第二级翻译:主语言 → 派生语言
一级语言完成后,将其内容作为基础,进行派生语言的翻译。对应关系如下:
基于
English (US):基于
繁體中文 (台灣):4. 翻译同步与提交
待所有一级和二级语言翻译完成后,我们将 Weblate 上的更改提交至仓库。
为了方便理解,我们还提供了清晰的流程图:
graph LR; subgraph s1["一级语言 (主语言)"] E2["日本語 (日本)"] E3["한국어 (대한민국)"] E4["繁體中文 (台灣)"] E1["English (US)"] E5["文言 (華夏)"] E6["梗体中文 (天朝)"] E7["焱暒妏"] end subgraph s2["二级语言 (派生语言)"] F7["Ελληνικά (Ελλάδα)"] F1["English (United Kingdom)"] F2["Español (España)"] F3["Français (France)"] F4["Русский (Россия)"] F5["Slovenčina (Slovensko)"] F6["繁體中文 (香港特別行政區)"] end D1["简体中文 (中国大陆)"] --> E1 & E2 & E3 & E5 & E6 D1 -- 转换器 --> E7 D1 -- 繁化姬 --> E4 E1 -- 转换用词标点 --> F1 E1 --> F2 & F3 & F4 & F5 & F7 E4 -- 繁化姬 --> F6 s1 ==> n1["最终译文"] s2 ==> n1单字符串翻译流程
此前介绍的流程侧重于整体语言的翻译方式,未涵盖具体的字符串翻译方法。本节将详细说明如何逐步翻译和处理单个字符串,以提高翻译效率与质量。
一、提交翻译建议
适用于尚未翻译或尚未核准的字符串。
识别源语言:明确该字符串所属的源语言,并参考上一节“翻译工作流程”,确定其所依赖的上一级语言。
初步翻译:使用 AI 工具 (推荐 ChatGPT 系列) 或翻译器完成初步翻译,确保内容通顺、语义准确,同时提高效率。
人工润色:基于初步翻译对文本进行优化,确保与原文保持一致,避免意译过度或误解原意。
长度评估与优化:在 PCL 界面中查找该字符串,预估翻译后的长度是否适合界面显示。必要时简化表达、删减冗词。
提交建议:确认翻译无误后,提交翻译建议。
graph LR A1[识别源语言] --> A2[AI 初步翻译] A1 -- 已初步翻译 --> A3 A2 --> A3[人工润色] A3 --> A4[长度评估与优化] A4 -- 长度不合适 --> A3 A4 --> A5[提交翻译建议]二、优化已翻译字符串
适用于已经核准但存在错误或可优化空间的字符串。
按照“提交翻译建议”的方式提交更新后的翻译内容。
发布评论以说明更改,评论类型选择
译文评论,与其他翻译人员讨论,格式如下:保存评论后,等待管理员审核。建议同步在 QQ 群或 Discord 中发起讨论以加快响应。
三、审核翻译建议
适用于已被翻译但尚未被核准的字符串。
核查建议内容:逐条查看建议,确保语义清晰、语法准确、符合上下文。
辅助工具校对:可使用 AI 工具辅助检查语法和语言习惯。
长度评估:在 PCL 界面中确认译文长度合适,如有冗长可适当优化。
执行核准:将
审校状态设置为已核准,点击保存。graph LR C1[检查翻译建议] --> C2{语法校对} C2 -- 语法错误 --> C5[接受并编辑] C2 -- 无问题 --> C3{长度评估} C3 -- 不合适 --> C5[接受并编辑] C3 -- 合适 --> C4[设为“已核准”并保存]四、重新审核已核准内容
适用于已核准但出现错误或存在优化可能的字符串。
查阅评论:阅读其他贡献者提交的评论,理解问题描述,若有疑问可在原评论下进一步询问。
判断合理性:评估建议是否成立。如不成立,拒绝建议并说明理由,标记评论为
解决。需要修改时:将
审校状态设为需要编辑,并保存,进入待修改状态。协商一致意见:在评论区或交流群中征求其他译者意见,达成共识后,重新进行翻译并走“审核翻译建议”流程。如果未达成一致意见,则拒绝建议并标记为
解决。graph LR D1[阅读评论并理解建议] --> D2{是否合理} D2 -- 合理 --> D3[设为“需要编辑”并保存] D3 --> D4{达成一致意见} D4 -- 未达成 --> D6[拒绝建议并标记为“解决”] D4 -- 达成 --> D5[重新翻译并执行审核流程] D2 -- 不合理 --> D6[拒绝建议并标记为“解决”]特殊语言翻译
一些语言的翻译步骤与其他语言略有不同,以下是一些特殊语言的翻译说明:
English (United Kingdom)
在将美式英语翻译为英式英语时,常规的翻译器或 AI 工具可能无法精准处理拼写与用词差异。因此,我们建议结合专业的拼写检查工具或转换器进行校对,以确保译文符合英式英语的规范。
推荐工具包括:
繁體中文 (台灣)
普通的翻译器、AI 工具或繁简转换器通常只能完成基础的字形替换,难以精准处理词汇和语义上的差异。例如:
网络 → 網路设置 → 設定内存 → 記憶體此外,由于汉字简化过程中多个字形被合并为一个字,导致在反向转换时往往无法准确判断应还原为何种繁体字,产生歧义。例如:
髮、發 → 发復、複、覆 → 复臺、檯、颱 → 台这若再辅以简单的查找替换,更可能出现啼笑皆非的误译 (
比如“海記憶體知己”) 。基于以上问题,我们选用了 繁化姬 作为标准转换工具。它不仅能将简体中文准确转换为繁体中文,还能覆盖大部分用词、用语差异,最大程度满足实际使用需求。
在转换后,手动校对并修改一些词就可以了。
使用方法
繁化姬 的操作非常简便,只需将简体中文粘贴至输入框并点击转换按钮,即可快速生成对应的繁体中文文本。推荐按以下步骤进行操作:
使用任意浏览器访问 https://zhconvert.org/。
切换至
詞語模組选项卡。开启以下推荐模组以提高转换准确性与覆盖范围:
修飾句子單位轉換專有名詞引號正、異體字重複字修正重複字通用修正電腦詞彙如有特殊需要,亦可根据情况调整开启的模组。但建议上述模组始终启用,以确保转换质量。
返回至
轉換选项卡。将简体中文文本粘贴至
轉換前文字输入框。点击
台灣化按钮。在
轉換後結果区域复制已转换的繁体中文文本。繁體中文 (香港特別行政區)
由于目前主流翻译工具普遍缺乏对
繁體中文 (香港特別行政區)的支持,AI 工具在处理香港用语时也常存在误差,因此我们同样推荐使用 繁化姬 进行转换。转换流程基本与
繁體中文 (台灣)相同,仅需注意以下两点:繁體中文 (台灣);香港化而非台灣化。由于繁化姬对香港本地用语的覆盖并不完全,建议在转换后手动校对并替换个别词汇,以确保最终译文更贴近香港本地表达。同时,请确保
正、異體字模组始终处于开启状态,以便输出符合港式书写标准的繁体字形。焱暒妏
火星文的翻译需要使用专门的转换器来完成,其本质上并非一种独立的语言,而是通过对字形的改变来实现特殊的视觉效果。
推荐使用以下工具进行火星文转换:火星文生成器。只需将原始文本粘贴到输入框中,即可快速生成对应的火星文文本。
文言 (華夏)
咕咕咕……
梗体中文 (天朝)
根据量子语义叠加原理与非线性符号动力学机制,在经典汉字拓扑结构的基底上,通过多维模因神经网络的量子化嵌入,梗体中文实现了对传统语言矩阵的亚稳态重构。具体而言,该语言变体在满足傅里叶语义守恒定律的前提下,通过高频模因共振因子与离散化梗元胞自动机的协同演化,在非平衡态语用学框架内形成了具有分形混沌特征的新型语义超结构,其核心参数化过程表现为对原始中文基底进行二次谐波振荡的语素重组,并经由互联网文化熵增通道注入具有自相似特征的亚文化符号素。顾名思义,梗体中文在原本的中文基础上,加入了大量网络流行语和梗元素()在实际翻译过程中,请尽情发挥你的想象力和创造力,集幽默与抽象于一身,
发挥美丽的夹子之力口巴!!!对于翻译的内容不会有太多限制,保持不过分不合适即可。PCL i18n & l10n Project Guide
This document is intended to help you quickly understand the PCL i18n Project workflow and point out common translation issues to contribute more efficiently and accurately. Whether you're a newcomer to translation projects or an experienced translator, this guide provides clear operational steps and practical advice. By following the processes outlined here, you can ensure the quality and consistency of your translations while avoiding common errors and misunderstandings. We hope this guide becomes a valuable assistant in your contributions to the PCL i18n Project!
Overall Translation Workflow
To ensure the quality and efficiency of the translation process, we have established the following workflow that balances accuracy, adaptability, consistency, and conciseness. Considering that directly translating from Chinese into certain languages (such as Western European languages) may result in lengthy or poorly structured text, we prioritize translating into English first, refining the expression, and then using it as the base version for other languages.
Therefore, it is essential to understand and adhere to the established workflow during the translation process to avoid non-standard operations. For example: submitting translations in
English (United Kingdom)that are entirely different fromEnglish (US)is not acceptable; derived language translations should not be created directly from Chinese; and relying solely on basic simplified-traditional conversion tools for language transformation is not permitted.The workflow is as follows:
1. Upstream Repository Update & Language File Preparation
When the main project repository (tangge233/PCL2/tree/Language) is updated, we will perform the following actions:
2. First-Level Translation: Chinese → Primary Languages
Newly added or updated Chinese strings are first pushed to the Weblate platform for the initial round of translation. The target primary languages include:
3. Second-Level Translation: Primary Languages → Derived Languages
Once the first-level translations are complete, their content will serve as the base for translating into derived languages. The corresponding relationships are as follows:
Based on
English (US):Based on
繁體中文 (台灣):4. Translation Synchronization & Submission
Once all first-level and second-level translations are complete, the changes made on Weblate will be committed to the repository.
To aid in understanding, a clear flowchart is provided:
graph LR; subgraph s1["Primary Languages"] E2["日本語 (日本) "] E3["한국어 (대한민국) "] E4["繁體中文 (台灣) "] E1["English (US)"] E5["文言 (華夏) "] E6["梗体中文 (天朝) "] E7["焱暒妏"] end subgraph s2["Derived Languages"] F7["Ελληνικά (Ελλάδα)"] F1["English (United Kingdom)"] F2["Español (España)"] F3["Français (France)"] F4["Русский (Россия)"] F5["Slovenčina (Slovensko)"] F6["繁體中文 (香港特別行政區) "] end D1["简体中文 (中国大陆) "] --> E1 & E2 & E3 & E5 & E6 D1 -- Converter --> E7 D1 -- ZhConvert --> E4 E1 -- Localization --> F1 E1 --> F2 & F3 & F4 & F5 & F7 E4 -- ZhConvert --> F6 s1 ==> n1["Final Translations"] s2 ==> n1Single String Translation Workflow
The previously introduced workflow focused on translating entire languages. This section provides a detailed explanation of how to translate and handle individual strings step by step, aiming to improve both the efficiency and quality of translations.
1. Submitting Translation Suggestions
This process applies to strings that have not yet been translated or approved.
Identify the Source Language: Determine the language the string originates from, and refer to the “Translation Workflow” section to find its corresponding parent language.
Initial Translation: Use AI tools (ChatGPT series recommended) or translation software to create a first draft. Ensure the translation is fluent and semantically accurate while maintaining efficiency.
Manual Polishing: Refine the draft translation to ensure consistency with the original text and avoid over-interpretation or misrepresentation.
Length Evaluation & Optimization: Locate the string in the PCL interface and assess whether the translated content fits well within the UI constraints. If needed, simplify expressions and remove redundant words.
Submit Suggestion: Once the translation is accurate and appropriately formatted, submit it as a suggestion.
graph LR A1[Identify Source Language] --> A2[AI Draft Translation] A1 -- Already Translated --> A3 A2 --> A3[Manual Polishing] A3 --> A4[Length Optimization] A4 -- Too Long --> A3 A4 --> A5[Submit Suggestion]2. Optimizing Approved Strings
This process applies to strings that have already been approved but contain errors or have room for improvement.
Submit the updated translation using the same procedure described in “Submitting Translation Suggestions.”
Post a comment to explain the change. Choose the comment type
Translation Comment: Discuss with other translators, and follow this format:After saving the comment, wait for admin review. To expedite the process, it is recommended to initiate a discussion in the QQ group or Discord channel.
3. Reviewing Translation Suggestions
This applies to strings that have been translated but not yet approved.
Review Suggested Translations: Carefully examine each suggestion to ensure the meaning is clear, the grammar is correct, and the translation fits the context.
Use Proofing Tools: Utilize AI tools to assist with grammar checks and to ensure the translation is aligned with natural language usage.
Length Evaluation: Check the translation within the PCL interface to confirm it fits appropriately. If it's too long, simplify the wording as needed.
Approve Translation: Set the
Review StatustoApprovedand click save.graph LR C1[Review Suggestions] --> C2{Grammar Check} C2 -- Grammar Issues --> C5[Accept & Edit] C2 -- No Issues --> C3{Length Evaluation} C3 -- Too Long --> C5[Accept & Edit] C3 -- Appropriate --> C4[Set as 'Approved' and Save]4. Re-reviewing Approved Translations
This process applies to translations that have already been approved but are found to contain errors or have potential for improvement.
Read Comments: Review comments submitted by other contributors to understand the reported issue. If anything is unclear, you may ask follow-up questions directly under the original comment.
Evaluate Validity: Assess whether the suggestion is reasonable. If not, reject the suggestion, provide an explanation, and mark the comment as
Resolved.If Edits Are Needed: Change the
Review StatustoNeeds Editingand save it. The string will then enter a pending modification state.Reach a Consensus: Discuss the issue in the comments section or through community channels (e.g. group chat) to reach a consensus. If agreement is achieved, revise the translation and proceed through the “Reviewing Translation Suggestions” process again. If no consensus is reached, reject the suggestion and mark the comment as
Resolved.graph LR D1[Read Comments & Understand Suggestion] --> D2{Reasonable?} D2 -- Yes --> D3[Mark as “Needs Editing” and Save] D3 --> D4{Consensus Reached?} D4 -- No --> D6[Reject Suggestion & Mark as “Resolved”] D4 -- Yes --> D5[Re-translate & Run Review Process] D2 -- No --> D6Special Language Translations
Some languages require slightly different translation procedures compared to others. Below are specific guidelines for handling special languages:
English (United Kingdom)
When converting American English to British English, standard translation tools or AI services may not accurately handle differences in spelling and vocabulary. Therefore, we recommend using professional spelling checkers or converters to proofread the translations and ensure they conform to British English conventions.
Recommended tools include:
繁體中文 (台灣)
Standard translators, AI tools, or simplified-traditional converters typically only perform basic character shape conversions and often fail to accurately address vocabulary and semantic differences. For example:
网络 → 網路设置 → 設定内存 → 記憶體Additionally, due to the simplification process where multiple traditional characters were merged into a single simplified form, it's often difficult to reverse-convert them correctly, resulting in ambiguity. Examples include:
髮, 發 → 发復, 複, 覆 → 复臺, 檯, 颱 → 台If simple find-and-replace tools are used on top of this, it can lead to laughable misinterpretations (
such as “海記憶體知己”).To address these issues, we have selected ZhConvert (繁化姬) as our standard conversion tool. It not only converts Simplified Chinese to Traditional Chinese with high accuracy but also handles most of the differences in vocabulary and usage, meeting practical translation needs to the greatest extent possible.
After conversion, a quick manual review and correction of a few terms will suffice.
How to Use
ZhConvert (繁化姬) is extremely easy to use. Simply paste the Simplified Chinese text into the input box and click the convert button to quickly generate the corresponding Traditional Chinese output. We recommend following these steps:
Open any web browser and go to https://zhconvert.org/.
Switch to the
詞語模組(Word Module) tab.Enable the following recommended modules to improve conversion accuracy and coverage:
修飾句子(Sentence Enhancement)單位轉換(Unit Conversion)專有名詞(Proper Nouns)引號(Quotation Marks)正、異體字(Standard & Variant Characters)重複字修正(Repeated Characters Correction)重複字通用修正(General Repeated Characters Fix)電腦詞彙(IT Terminology)If needed, you may adjust the enabled modules based on the specific context, but it is strongly recommended to keep the above modules active for optimal quality.
Return to the
轉換(Convert) tab.Paste the Simplified Chinese text into the
轉換前文字(Text Before Conversion) input box.Click the
台灣化(Taiwanize) button.Copy the converted Traditional Chinese text from the
轉換後結果(Converted Result) section.繁體中文 (香港特別行政區)
Due to the limited support for
繁體中文 (香港特別行政區)in most mainstream translation tools, AI models also often struggle with accurately handling Hong Kong-specific expressions. Therefore, we similarly recommend using ZhConvert (繁化姬) for conversions.The conversion process is nearly identical to that for
繁體中文 (台灣), with just two key differences to note:繁體中文 (台灣)香港化(Hong-Kongize) button instead of台灣化(Taiwanize).Since ZhConvert does not fully cover all nuances of Hong Kong-specific terminology, it is recommended to manually proofread and replace individual words after conversion to ensure that the final translation better reflects local usage. Additionally, be sure to always enable the
正、異體字(Standard & Variant Characters) module, so the output conforms to the traditional character forms used in Hong Kong writing conventions.焱暒妏
The translation of
焱暒妏(Martian Script) requires a specialized converter. It is not an independent language in the traditional sense, but rather a stylized form of writing achieved by altering character shapes to create a unique visual effect.We recommend using the following tool for Martian script conversion: Martian Text Generator. Simply paste the original text into the input box, and the tool will instantly generate the corresponding Martian-style text.
文言 (華夏)
WIP……
梗体中文 (天朝)
Based on the principle of quantum semantic superposition and the mechanism of non-linear symbolic dynamics, Meme-style Chinese reconstructs the metastable structure of traditional language matrices atop the topological foundation of classical Chinese characters. Specifically, this linguistic variant, under the constraint of the Fourier Semantic Conservation Law, evolves through the synergistic interaction of high-frequency meme resonance factors and discretized meme-cellular automata. Within a nonequilibrium pragmatics framework, it gives rise to a novel semantic superstructure characterized by fractal chaos. The core process involves a second-order harmonic oscillation of morphemes over the original Chinese base, injecting self-similar subcultural symbols through the entropy-expanding channel of internet culture.As the name suggests,
梗体中文 (天朝)(Meme-style Chinese) is based on standard Chinese but infused with a rich mix of internet slang and meme elements. During the actual translation process, feel free to unleash your imagination and creativity—blend humor with abstraction, and don’t hold back from embracing the ✨chaotic energy✨ of meme culture.Let your inner gremlin shine through, yasss!!!There are very few restrictions on what can be done in Meme-style translations—as long as the result isn’t excessively inappropriate, you’re encouraged to have fun and be as playful and expressive as you like.
附录 2:.NET 中的本地化 from Microsoft
Beta Was this translation helpful? Give feedback.
All reactions