コンテンツ翻訳/説明文書/よくある質問

This page is a translated version of the page Content translation/Documentation/FAQ and the translation is 17% complete.

これはコンテンツ翻訳についてよくある質問集です。

コンテンツ翻訳って何？

他言語版に同じトピックがあり、編集者がそれについて自分の言語で新しい記事を書きたい時に補助する機能で、ツールのセットと仕様を提供します。

ウィキペディアの全言語版に導入されたのは2015年です。

CXと呼ぶ理由は？「CT」じゃないの？

「CX」は英語「Content Translation」（コンテンツ翻訳）の略号ですが工夫してあります。実は「CT」を使いたくても、すでに拡張機能のCategoryTree をその略称で呼んでいるため、別の略号を選びました。

英語で書くと正しいのは「ContentTranslation」か「Content Translation」か?

「ContentTranslation」が、このプロジェクトを可能にする MediaWiki 拡張機能の技術名です。

For everything else it's preferable to write "Content Translation", with a space.

The name can and should be translated to other languages, as separate words, and capitalized according to the respective languages' rules.

既存の記事の翻訳に、コンテンツ翻訳ツールを使えますか?

No.

It is possible to publish to the same title. You will see a warning if you try to do this. If you publish despite the warning, and ContentTranslation will create a new revision. Old revisions will be preserved in the history, but your new revision won't have any of their text.

Section Translation is a new feature in development, and it will allow to translate just one section and add it to an existing page without discarding the other existing sections.

CX2って何?

It's a major version of Content Translation first released in September 2018. It keeps the general design of Content Translation, but several features are thoroughly updated. In particular, it integrates VisualEditor as the editing environment, improves paragraph alignment, and makes several changes in support for templates. As of October 2018, CX2 is still in development. You can try it by clicking "Try the new version" on the Content Translation dashboard.

The version before CX2 is called "CX1".

For more information about CX2, see コンテンツ翻訳/バージョン2 .

CX3って何?

It's a major version of Content Translation, which is being developed during 2020. It is also sometimes known as "the Boost project", "Section Translation", and "SX". It will include better adaptation to mobile devices, the capability to translate sections rather than whole articles, new entry points, and other features. For more information, see コンテンツ翻訳/節単位の翻訳 .

このツーつの使い方マニュアルはないの？

Yes! See the page Content translation user guide. You can also translate it to your language.

コンテンツ翻訳ツールは機械翻訳とか自動翻訳と同じ?

Neither.

Content Translation is a working environment for human wiki editors that helps creating the first revision of an article by translating an existing article from another language. It automatically helps to adapt images, links, footnotes, templates and some other formatting elements.

For some language pairs it also has the option of integrating a machine translation service. The currently supported services are Apertium, Google Translate, Matxin, Yandex.Translate, and Youdao. More services may be added in the future.

どれくらい人気がある?

Definitely! It was pretty clear even before the development started: In the past there were so many attempts at making similar tools that it's impossible to count them. Some are listed at m:Machine translation#Attempts (please add there any you know of).

As of October 2018, over 40,000 users created over 370,000 articles using it since it was enabled in January 2015. More than 500 users created over 100 articles each using Content Translation. So it is more certain than ever that there is demand for it.

翻訳拡張機能とコンテンツ翻訳ツールの違いは何?

The Translate extension was initially built with focus on translating software user interface messages for MediaWiki and other programs. As such, it's built to keep translations in sync with the source language, while each language edition of a Wikimedia project is supposed to be independent and not reflect a source language.

To provide its advanced features on MediaWiki wiki pages, Translate requires preparation of the source page for translation with some additional markup, which can't yet be handled visually. Said wikitext is distracting to most editors and we did not want to expose Content Translation translators to it.

同じウィキなら誰でも使えますか?

It is available for logged-in users of a wiki where it's enabled, and it must be enabled as a beta feature in the preferences.

It is gradually being taken out of "beta feature" status, and converted to a user preference.

コンテンツ翻訳ツールを試そうとするとページは白紙のままで何もできません。解決策は?

The most common reason for this is old local JavaScript code on your wiki, which is incompatible with Content Translation. Such code usually appears in Gadgets, user scripts, or Common.js. If you recently enabled a gadget or changed a user script, try disabling it. If this fixes the problem, talk to the gadget maintainers on your wiki and ask them to fix it.

The incompatibility is usually resolved by going to Special:CX and finding the failing source code line at your browser's JavaScript console. Most often the reason for the failure is that the #bodyContent element is used in the JavaScript code. This element is not available on Special:CX, you should edit the JavaScript code and check whether #bodyContent actually exists.

If this explanation is too complicated or doesn't help, please report the issue at Talk:CX, and if possible, notify the people who maintain gadgets and user scripts on your wiki.

言語間リンクがグレーい反転しているなので、どうすれば黒色に変わりますか?

If you have Content Translation enabled in your user preferences or as a beta feature, Content Translation will sometimes add up to three gray interlanguage links to Wikipedia articles. A gray link is shown if the article is not available in a relevant target language. Relevant target languages (no more than three) are selected for each user in the following order:

The user interface language, as selected in the preferences (it may be different from the wiki's content language).
The user interface language of the browser's menus.
The user's preferred content languages, as defined in the browser preferences.
Languages on which the user previously clicked in the language selection panel.

ウィキペディアの記事以外は翻訳できないの ?

The focus for initial development is articles in Wikipedia. In the future it may also be enabled in Wikivoyage and other sites.

コンテンツ翻訳ツールで新しい記事を作成する手順は?

The main entry point to Content Translation is a button on your contributions page:

Click "投稿記録" in your personal bar (near "ログアウト").
Click "新規執筆" and select "翻訳".
Click "新しい翻訳".
Select the language from which you want to translate in the "From:" field, and type the name of the article in that language.
Select the language to which you want to translate and type how the new article will be called.
Click "翻訳を開始". This will take you to the translation interface.
Type the translation of each paragraph in the translation column. You don't have to translate all the paragraphs. Translate as much as needed for the wiki in your language.
Until you publish, the translation is regularly saved automatically, so you don't have to worry that you'll lose it. To come back to an article that you started translating, repeat steps 1 and 2 and select the article from the list that you'll see.
When you wrote everything you want for the first version of the new translated article, click "翻訳を公開……". This will create a new article in the wiki in the target language with the title that is given at the top of the translation column.

What are the "Suggestions" in the dashboard?

"Suggestions" are personalized automatically-generated links to translating articles that exist in one language but not in another. They are generated based on the articles that the user who is viewing the CX dashboard translated recently, and on frequently-read articles in one language that don't exist in other languages. So for example if you translated an article about a city in India, you are likely to see more suggestions to translate more articles about cities in India. 当選直後のスロバキア大統領について、英語版には記事があるのに日本語版にない場合は、翻訳のお勧めが表示されるかもしれません。

翻訳前に指定の記事を原語で読むには、どうすれば？

言語名をクリック。

一例として指定では英語原文をイタリア語に翻訳するよう求められた場合なら、指定欄の最下部に「English > Italiano」という記述があります。「English」という文字列は英語版へのリンクのはずです。

コンテンツ翻訳機能はリンクをどう処理する？

システムは原文内のリンクに言語間リンクがないか検索し、翻訳先の言語にリンクが見つかったら自動で訳文内に挿入します。

リンク追加には次のような方法もあります。

訳文内で単語を選択するときは原文のつづり（スペル）のままに記して、サイドバーの「リンクを追加」ボタンを押します。
訳文内の単語を選択してから、原文欄のリンクをクリック。 If a corresponding article exists in the target language, a link to it will be added.
Selecting a word in the translation and adding a link to any page using the tool that appears in the sidebar. By default, a list of available adaptable will be shown under the link tool.

How does Content Translation handle categories? Can I add categories or use HotCat?

Content Translation automatically adapts categories for which there is a directly-corresponding category in the target language. The correspondence is derived through interlanguage links.

If you use CX1, if you want to add more categories to an article, it will have to be done after the initial version is created.

If you use CX2, you can add more categories during translation by typing their name at the bottom of the categories list in the translation column.

How does Content Translation handle references?

The tool will try to adapt references as much as possible between the source and target languages.

If you delete a reference from the translation, you can add it back by placing the caret where you want to add the reference, clicking the reference in the source column and then clicking "Add reference".

Adapting references may be challenging given that different languages use different citation formats. If there is a reference that you cannot adapt, or that is adapted incorrectly, please report a bug.

Is it possible to edit references in Content Translation?

Adding references is not possible in CX1. You can only to copy them as-is. You can change them or add your own references after creating the article.

In CX2 you can add, edit, and delete references similarly to how it is done in Visual Editor.

翻訳原文から画像も転写できませんか?

Yes, images will be copied just like paragraphs - simply by clicking. The translator will have to type the caption, of course.

It works only if the image is stored in a common media repository (for Wikimedia projects this is Commons). It doesn't work for files stored in the projects locally. It also won't work if the image is a part of a template, such as an infobox; however, it may be possible to adapt the image if the template is adapted.

記事を公開するって何?

The publishing of the article is the same as creating a revision of a wiki page - it has a date, the revision author's username and an edit summary, it appears in Recent Changes, etc.

How are the writers of the source article credited?

The creators of the source article are credited in a link to the revision of the source article from which the translation was made, which is added automatically to the edit summary of the published version. This is compatible with the CC BY-SA license and with Wikimedia's Terms of Use, which require you as the re-user to give attribution "through hyperlink [...] to the page or pages that you are re-using".

Will the previous revisions of the translated be imported when I am translating?

Not automatically.

In Wikipedia in some languages, such as German, there is a custom of importing the revisions of the article that was translated so that they appear in the history as if they were made before the first actually translated revision. This can be done with articles that were created using Content Translation in the same way.

Can I continue translating after publishing?

Phabricator で追跡中
タスク T102966

At the moment Content Translation is focused on creating the first version of the article. After publishing, the translation cannot be loaded as such from the dashboard, and the published page must be edited as a usual wiki page.

There is a plan to make it possible to add translated paragraphs to already-published pages.

Will Content Translation use information from Wikidata?

Yes, in several ways:

Interlanguage links from Wikidata are used to auto-fill the links in the translated article.
When a translated article is published, a sitelink to it is added to the corresponding Wikidata item, so that an interlanguage to it appears immediately.
Descriptions from Wikidata are shown in the linking tool and in the article selection tool.
The dashboard has a link to a Wikidata-based tool that helps you find articles that don't exist in your language.

When templates in different Wikipedias will use data from Wikidata more, they will be simply picked up by Content Translation without any effort from the translator.

There are also plans to use labels and aliases in smart ways in the future.

What are the translation aids that will be made available?

The current plan is:

Dictionaries – translation and definitions of words.
Link adaptation – Links will be adapted automatically when they will be available as interlanguage links to the target languages. It will be possible to make basic manipulation on them – remove them and pick them from other sources.
Category adaptation – Categories that have a directly corresponding category page in the target language linked by an interlanguage link will be added to the translated page.
Image adaptation – Images are copied to the translated article in one click.
Machine translation and translation memory – These are similar to what is used in the Translate extension.
Automatic interlanguage link adding–- An interlanguage link will appear immediately after publishing the article.

Can anybody read the text that is saved automatically while I am writing the translation?

Phabricator で追跡中
タスク T90163

No. It is accessible only to the translator until the publishing.

Will you provide suggestions from translation memory?

This is planned as a future feature.

The data for translation memory will have to be filled from some initial translations, so it may take a while from the time that translation memory is enabled for Content Translation until it becomes useful.

How good are the articles created using Content Translation?

As good as any other articles are created in Wikipedias in the respective languages.

Since the deployment of Content Translation as a beta feature in some languages in January 2015 until November 2015 about 30,000 articles were created. In July 2015 CX became enabled as a beta in all languages, and since then till November 2015 less than 10% of the articles created using CX were deleted. For comparison, the rate of deletion of articles that are created using the wiki syntax editor goes up to 50% in English.

The articles that were not deleted developed as usual Wikipedia articles: people fixed layout, added or edited paragraphs, added templates, improved references, and so on.

Usually these improvements were done both by the person who created the first version and by other wikipedians.

There is no machine translation for my language. How is Content Translation useful to me and my wiki?

By itself Content Translation is not a machine translation tool. Its primary focus is to help people to create translated wiki pages as efficiently as possible. It includes tools that are tightly integrated with MediaWiki and its usual content creation and editing workflow: display of the source and the translation side-by-side; adaptation of links, categories, images and text formatting; publishing to different namespaces; interlanguage links These features are already supposed to make typing translated articles by hand easier.

This is not just theory. Content Translation was enabled in the French Wikipedia on March 31 2015 and by June 7 it was used to create 500 articles, even though machine translation was not available.

The fact is that machine translation is not available for the majority of languages in which there are Wikipedias, so most language pairs will only be able to use Content Translation as a tool to translate articles manually with the above adaptation tools. If you want to help create a machine translation engine for your language, see How can I improve machine translation support for my language?

Machine translation to my language is bad, and it's easier to translate manually. How is Content Translation useful to me and my wiki?

As written in the previous answer, Content Translation is not by itself a machine translation tool, but a tool to create translated wiki pages. It is designed to be useful even without machine translation.

Machine translation works quite well in some languages, and then it can make the translators' work even more efficient. Machine translation support for a language pair is enabled after testing and approval from people who know the language well.

If machine translation support for your language is enabled, but you don't want to use it, you can disable it and still enjoy the other tools, such as link, category, and image adaptation, as well as dictionaries (if available for your language).

機械翻訳をどうやって取り入れますか?

For language in which machine translation is supported in Content Translation, machine translation will be auto-filled upon clicking a paragraph in the translation area.

Initially we're using the Apertium engine, which is free software and can be installed and maintained on our own servers. At a later point we may use Moses and other engines. In November 2015 we added Yandex for limited use between English to Russian and in the months after that we enhanced this to most languages that Yandex supports. We have added more machine translation services since then - Matxin and Youdao. 手順の詳細はコンテンツ翻訳と機械翻訳システムをとりあげたページをご参照ください。

CX2 が原文から単に文字列をコピペします。内容をきちんと翻訳させたいのですが？

First of all, it's important to understand that Content Translation by itself is not a machine translation tool. The translation is supposed to be done by humans. For some languages machine translation is integrated, but even then it is always supposed to be fixed by the translator.

If Content Translation supports machine translation to your language, you'll be able to enable it in the "Automatic translation" card that appears in the sidebar when you click a paragraph.

If machine translation to your language is not supported, you will only have two options: to retype the information that is initially pasted in the source language, or to disable it entirely in the "Automatic translation" card and type everything by yourself. In all cases you'll be able to use other tools, such as link, image and reference adaptation.

Yandex はどの言語を処理できますか？増やさないのですか？

Yandex is available at present for more than 70 languages. As Yandex’s language coverage expands we will consider enabling them for Content Translation. Please note: Yandex machine translation will not be available when creating pages into English.

Yandex と Apertium では使い方などに違いはありますか？

As a user of Content Translation you will not feel any difference on the translation interface as the machine translation system of Yandex will display the translated content in the same way Apertium currently does for the supported 45 language pairs.

Yandexを使用すると機械翻訳の作業はどう進みますか?

Yandex provides a free for use API key that allows websites and other services to use their translation system. Content Translation also uses a unique API key to access this service on Yandex's server. When a user starts translating an article, the HTML content of each section of the source article is sent to the Yandex server and a translated version is obtained and displayed on the respective translation column of Content Translation. Links and references are adapted as usual and users can modify the content as required.

This process continues for all the sections of the article being translated. For better performance, the translations for consecutive sections are pre-fetched. The user can save the unpublished translation (to work on it again at a later time) or publish the article in the usual manner. The article is published on Wikipedia like any other normal article with appropriate attribution and licenses.

You can view a diagram of the process.

Yandex はオープンソースのソフトウェアじゃないはず。それなのに使って良いの？

Content Translation evolved from a long-standing need to bridge the gap in the amount of content between Wikipedias in different languages. Like all other software used on Wikimedia sites, Content Translation is also open source. In this particular case as well, we are using an open source client to interact with the external service and import freely licensed content in order to help users expand our free knowledge.

To use Yandex’s machine translation system we are not adding any proprietary software in the Content Translation code, or on the Wikimedia websites and servers. The service is free of charge and available for everyone.

Only the freely available Wikipedia article content (in segments) is sent to the Yandex service and the obtained translated content is freely usable on Wikipedia pages. The translated content can be modified by users and this data is also available publicly under a free license through the Content Translation API. This is a valuable resource made available for the community to develop open source translation services for those languages where they don't exist yet.

After studying the implications carefully, we found the fact that the content was stored previously in a closed source service does not limit the freedom of our knowledge or our software in the present or the future. We have taken special care to make sure that the content provided is freely licensed to make sure it complies with Wikipedia policies. This includes a long process for legal and technical evaluation and compliance. The summary of the terms of use is also available.

From user feedback we have seen that machine translation support is really helpful for users and we want to support all languages in the best way. Guided by the principles of Wikimedia Foundation’s resolution to support free and open source software, we will prioritise the integration of open source services whenever they are available for a language. Apertium has been a critical part of Content Translation since its inception, but currently it only provides machine translations for 45 of the numerous possible language combination that Wikipedia can support.

Yandex を使うと個人情報は大丈夫？

Irrespective of the service being used, you can be sure that only Wikipedia content from existing articles is sent and only freely licensed content will be added back to the translation. No personal information is sent and communication with those services happen at the server side, so they are isolated from the user device. Please refer to this diagram for more details.

もしも機械翻訳が Yandex しかなくて使いたくないときはどうすれば？

Machine Translation is an optional feature in Content Translation that you can easily disable at will. If more machine translation systems are added for your languages, you can choose to enable MT again and select the MT service of your choice.

コンテンツを Yandex で翻訳した結果はウィキペディアで無償で使えるの？

Yes. The content received from Yandex is otherwise freely available on the Yandex web translation platform. Content Translation receives it via an API key to make it seamlessly available on the translation interface. This content can be modified by the users (if necessary) and used in Wikipedia articles under free licenses.

この内容は機械翻訳全般の改良に利用できますか？

Yes. Translations made in Content Translation are saved in our database. This information will be made publicly available for anyone to use as translation examples to improve their translation services (from University research groups, open source projects to commercial companies, anyone!). The content can be accessed via the Content Translation API. Please note, only information related to translated text is publicly available. This includes - source and translated text, source and target language information and an identifier for the segment of text.

私の言語で機械翻訳を改善したいのですが、何をすれば?

Apertium

Apertium is an open source rule-based machine translation engine. If you know your language well and you know programming you can try adding support for your language to Apertium.

Contribute to an existing Apertium pair, or create a new one!

Get in contact with the Apertium community with IRC, #apertium ^接続, or many other ways.

統計型機械翻訳エンジン

Statistical machine translation engines such as Google, Yandex, and others, need a lot of texts in different languages. If your language is poorly supported in these engines, or not supported in them at all, the best thing that you can do to contribute to the development of machine translation into your language is to translate as many articles as possible into it. This will take effort and time, but that's how it works.

Why doesn't Content Translation use the wiki syntax editor?

Because it should be easier for translators who are beginners with Wikipedia editing, and because it was much easier to implement features the adaptation of references, images and links, as well as machine translation integration in an HTML-based WYSIWYG editor. Content Translation is an article creation tool rather than an article editing tool. Because it is not supposed to be a full-fledged article editing environment, it only provides the most basic formatting tools. After an article is created, it can be edited in the VisualEditor or in the source editor, just like any other article.

In more technical terms, Content Translation uses a simple HTML "contenteditable" element that is available in modern browsers. It transforms the source article's HTML to the translation, and when publishing the article as a wiki page, it converts the translation to wikitext using Parsoid. At the moment, Content Translation does not use the VisualEditor for editing the translation, though this may be done in the future.

As of mid-2016, there is no plan to add the ability to write the translation in wiki syntax in Content Translation. There is a possibility that this will happen in the future, however, for example by integrating the 2017 wikitext editor that was developed by the Visual Editor team.

If you are concerned about the quality of the wiki syntax, you can publish your translation to user or draft space by changing the target title appropriately.

Are you building on other efforts as well?

There was a lot of research on the topic, see Machine translation#Attempts. For instance: «The quantitative results show that the contributions can improve the accuracy of a combination of RBMT-SPE pipeline at around 10 %, after the post-edition of 50,000 words in the Computer Science domain. We believe that these conclusions can be extended to MT engines involving other less-resourced languages lacking big parallel corpora or frequently updated lexical knowledge». (doi:10.1007/978-3-642-35085-6_4)

Can the machine-translated content be edited manually?

Yes, and it should be!

We treat machine translation only as a tool that may help a human translator be faster. Publishing machine-translated articles is not the intention of Content Translation, and it is actively discouraged.

Will there be a feature to prevent bulk publishing of unedited machine translated text?

Yes! We take article quality seriously. Machine translation is only a tool that helps the translator be more efficient, and the developers understand well that all translations must be edited by a human. The translation interface will show a warning if the translator will try to publish an article that only has machine translation. The developers will work with the editing communities to adjust this for the needs of every language.

What dictionaries will be available?

The dictionaries will be initially taken from free dictionaries from the freedict project. Later other dictionaries may be added, such as Wiktionary, OmegaWiki, terminology collections, and possibly other open sites.

How will templates be handled? How are you handling infoboxes?

Since November 2016 it is possible to edit templates in Content Translation. For each template, three options are given:

Skip the template.
Transfer the template's source code as-is to the translation.
If a corresponding template exists in the target language, insert that corresponding template and edit its parameters one by one.

CXのテンプレート対応は、コンテンツ翻訳/テンプレートをご参照ください。

Is it possible to insert a template using Content Translation?

If you use CX2, you can insert templates just like in Visual Editor.

It's not possible in CX1.

Will I be able to use the ULS input methods?

Yes!

ホームウィキの特定のページに翻訳ガイドラインへのリンクを張っても良いですか？

Yes. To do this, a user with appropriate permissions must edit the page MediaWiki:Cx-tools-view-guidelines-link.

私の言語でこれを使えるのはいつごろ?

All Wikipedias are supported since July 8, 2015.

ツールの技術面を詳しく書いてあるのはどこ？

Start from the following pages:

私のローカルのウィキでもコンテンツ翻訳拡張機能を設定できますか？

Yes. Just install the extension and follow the configuration guide. The default configuration has a bias for Wikipedia, so be sure to set it up correctly for your wiki.

cxserverとは？

The ContentTranslation extension works from the outset with multiple wikis and it needs to synchronize information between them. To make this possible, it uses an additional component called "cxserver". It also optimizes much of the connection to translation tools, such as dictionaries, machine translation, etc.

Microsoft Internet Explorerでも使えますか?

Content Translation has the same browser support policy as VisualEditor. It is supposed to work in browsers listed as "Modern / Grade A" on the page Compatibility. Microsoft Internet Explorer 11 and newer versions are supported. Don't expect it to work on versions 10 or older.

用語集

annotation: A markup applied to some part of text. Basically, it is html tags like anchor, bold, italic, underline etc.
card: a box which appears in the tools column on the special page and provides translation tools for specific context, e.g. a box that allows editing links
columns: vertical areas in which Special:ContentTranslation is divided: there are currently three columns (source, translation, tools)
Content Translation (CX): This tool consisting of ContentTranslation extension and cxserver backend.
cxserver: Backend for CX written in Node.js, handling text segmentation and providing consistent API for services like machine translation, dictionaries and translation memories.
glossary: A list of terms with definitions or translations.
GWT (Given-When-Then): GWT is a semi-structured way to write down test cases. They can either be tested manually or automated as browser tests with Selenium.
lemmatization: also called stemming. Mapping multiple grammatical variants of the same word to a root form; e.g. (swim, swims, swimming, swam, swum) -> swim. Derivational variants are not usually mapped to the same form (so happiness !-> happy).
link localization: Converting a wiki article link from one language to another language with the help of wikidata. Example: http://en.wikipedia.org/wiki/Sea becomes http://es.wikipedia.org/wiki/Mar
machine translation (MT): Initial translation made by computer algorithms to help translating faster.
morphological analysis: mapping words into morphemes, e.g. swims -> swim/3rdperson_present
parallel bilingual text: two versions of the same content, each written in a different language.
segmented: reduced in segments
segment: Smallest unit of text which is fairly self-contained grammatically. This usually means a sentence, a title, a phrase in a bulleted list, etc.
segmentation algorithm: rules to split a paragraph into segments. Weakly language-dependent (sensible default rules work quite well for many languages).
sentence alignment: matching corresponding sentences in parallel bilingual text. In general this is a many-many mapping, but it is approximately one-one if the texts are quite strict translations.
service: Things like MT, TM, Glossary
service providers: External systems which provide a service. Example: Google
source column: the column showing the segmented article in source language.
template destruction: inlining a template contents when suitable template does not exist in the target wiki
tools column: the column where cards appear
translation column: the column where the translation is done.
translation memory (TM): A service which suggests translations based on previous translations.
translation tools (translation support tools, translation aids): Context-aware translation tools like MT, Dictionary, link localization
word alignment: matching corresponding words in parallel bilingual text. This is strongly many-many.
translation dashboard: Listing of all translations of a user. A new translation can also be started from here.

コンテンツ翻訳ツールで「公開」を押すとフリーズする。これを予防するには？同じ記事で何度も発生しないように予防できないのか？

フリーズの回避

Force the browser to quit (There are no guarantee)

再発防止

Remove the text that seems to be causing the freeze.

At that time, cut the text using the visual editor of the content translation tool, and save it to another location using the usual visual editor.

As a precautionary measure, be sure to empty the entire paragraph using "最初に空行から開始します", rather than just removing the text.

Once published, add the text you saved earlier and make any necessary corrections.

「疑わしい」バグの解決策を書いてください。

ソースコード編集機能を使うには

コンテンツ翻訳ツールにはソースコード編集機能がありませんが、これはバグではなくこの機能の特性です。そのため、エラーの原因を特定しにくい場合もあります。現状では、例えば個人用の下書きページ（sandbox）でビジュアル編集機能を実行しておき、コンテンツ翻訳ツール内の編集機能で怪しいと感じた文をコピーして、先ほどの別ページ（sandbox）にペーストします（転写）。こうすると、下書きページでソースコード編集を使えます。

編集フィルタの挙動がコンテンツ翻訳で正常ではないのですが

Edit filters may behave improperly in content translation.

(An example and about its causes see Topic: Virkxk0741k38j8l, Topic: Vjsragp1ufrx9b18.)

Workaround: Remove paragraphs, etc. that were subject to errors, perform separate translation work, and add them after "publishing" the content translation.

利用者が訳文にカテゴリを追加したのに、公開すると表示しません

この問題はphab:T242613に報告済みです。

回避するには：コンテンツ翻訳で作業中に、テキスト編集機能で書き換えておき、「公開」した後に編集で追加する。