如何在进一步翻译中重复使用审阅过的文档？
我们的说明手册采用 Markdown 格式。我们每天或每周改进手册，然后我们想翻译成其他语言。到目前为止，这个过程是手动的，我们使用TRADOS（一种辅助翻译软件），
我们正在研究改进这个过程。我们想要做的是使用 Microsoft 翻译器翻译它，然后由人工审阅者进行微调。问题是我们如何在 Microsoft Translator 中重新使用纠正后的文档进行下一次翻译。
很多时候，我们只更改一个主题的 2 或 3 个单词，然后译者就会创建一个完整的新翻译，如果翻译不准确，审稿工作就会非常繁琐。
我知道我们可以训练模型，但我认为译者使用评论的可能性并不是 100%。而且，在审阅文档后维护失真似乎非常耗时。
我想知道是否有人解决了此类问题。
Our instruction manual is in markdown format. We improve the manual daily or weekly and then we would like to translate to other languages. Until now, this process is manual and we use TRADOS (a assisted translation software)
We are studying to improve the process. What we want to do is to translate it using Microsoft translator and then a human reviewer do the fine tune. The problem is how we can reuse the corrected document in Microsoft Translator for the next translation.
Many times we only change 2 or 3 words of a topic and then the translator would create a complete new translation, doing the reviewer work very tedious if the translation is not accurate.
I know that we can train the model, but I think that there is not a 100% of probabilities that the translator uses the review. Also, it seems very time consuming to maintain the distortionary after reviewing the document.
I was wondering if somebody has solved this kind a problem.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（3）
我们可以使用 Microsoft Flow 自动翻译文档。在Microsoft Flow中我们可以使用SharePoint服务来存储我们需要翻译的文件，然后我们可以创建一个操作流程。
注：不能保证翻译后的含义与原文相同。
We can automate the translation of the document using Microsoft Flow. In Microsoft flow we can use the SharePoint service to store the file which we need to translate and then we can create a flow of operation.
Note: Can't assure that the meaning after translation will be the same as the original statement.
您正在寻找翻译记忆库系统 (TMS)。 TMS 将存储您对文档的人工编辑，并将其重新应用到将来重复段落的文档翻译中。 TMS 还将帮助您的翻译人员找到新片段与之前翻译的片段的密切匹配，突出显示翻译人员必须执行的更改。
大多数 TMS 集成了 Microsoft Translator 等机器翻译系统，作为对人工编辑的建议，然后人工编辑可以批准或编辑该建议。
https://en.wikipedia.org/wiki/Translation_memory
选择 TMS 时，这里有一些需要考虑的功能：
一旦您收集了大量特定领域的人工翻译，您就可以训练自定义翻译系统直接从您的 TMS 内容中调整未来的机器翻译，以适应您的人工翻译所体现的术语和风格。
You are looking for a translation memory system (TMS). A TMS will store your human edits to documents and reapply them to future translations of documents where a paragraph is repeated. A TMS will also help your human translator find close matches of a new segment with a previously translated one, highlighting the change the human will have to perform.
Most TMSes integrate machine translation systems like Microsoft Translator and others as a suggestion to the human editor, who can then approve or edit the suggestion.
https://en.wikipedia.org/wiki/Translation_memory
When selecting a TMS here are some features to consider:
Once you have collected a significant amount of domain-specific human translations, you can train a custom translation system directly from the content of your TMS, tuning future machine translations in the direction of the terminology and style that are exemplified by your human translations.
您说您已经使用 Trados。那么，为什么不在 Trados 中创建翻译记忆库，并将更正的作品保存到该翻译记忆库中。然后，下次您在 Trados 中创建项目时，可能还会使用 MS Translator 插件，该插件可从 免费获取在 RWS AppStore 中，您将使用翻译记忆库和 MS Translator 进行预翻译。您之前翻译过且位于翻译记忆库中的任何作品都将优先于机器翻译结果使用。
如果您有专业版 Trados，您还可以使用 Perfect Match。如果您这样做，那么当您收到仅需要更改几个单词的更新文档时，您可以将其与之前翻译的双语文件进行匹配。保持不变的所有内容都会获得“完美匹配”状态，并且可以选择锁定，从而可以轻松识别译者需要更改的内容。
You said you already use Trados. So why don't you create a Translation Memory in Trados, and save your corrected work to that Translation Memory. Then the next time you create your project in Trados, presumably also using the MS Translator plugin that is freely available from the RWS AppStore you will pre-translate using your translation memory AND MS Translator. Any work you translated before and is in your translation memory will be used in preference to the machine translation result.
If you have the professional version of Trados you can also use Perfect Match. If you do this then when you receive your updated document with only a few words to change you match it against the bilingual file from your previous translation. Everything that remains the same receives a Perfect Match status and is optionally locked making it simple to identify what needs to be changed by the translator.