The Signpost

Op-ed

Content Translation beta is coming to the English Wikipedia

Contribute  —  
Share this
By Runa Bhattacharjee, Amir Aharoni, and WMF Language Engineering team
Screencast showing how to use Content Translation (demo starts at 0:40)

The Wikimedia Foundation's Language Engineering team plans to introduce Content Translation—a tool that makes it easier to translate Wikipedia articles into different languages—as a beta feature on the English Wikipedia.

Content Translation is an article creation tool that allows editors to quickly create an initial version of a new article by translating from an existing Wikipedia page of the same topic in another language. The tool is currently available on 224 Wikipedias as a beta feature, and more than 7,000 articles have been created by more than 1,500 editors since January 2015. For the English Wikipedia, we expect to enable as an opt-in beta feature for logged-in users in early July, after its initial Wikimedia testing in July 2014 and beta deployment to multiple Wikipedias in January 2015.

Content Translation saves translators' time by automating common tasks: it adapts formatting, images, links, categories and references, and it automatically adds an interlanguage link. The translated articles created are otherwise just like any other; the tool simply helps create quality content by allowing editors to focus on translating and expanding articles instead of being consumed with the lengthy manual translation process.

Once the tool is activated on the English Wikipedia, you will be enable to the Content Translation by going to your preferences, you can start a translation in the following ways:

Content Translation tool icon

These features will be visible only to the users who enable Content Translation. After doing this, you can start translating by following these steps:

  1. Open the tool by going to Special:ContentTranslation or to your contributions page.
  2. Click on the button to create a new translation.
  3. In the displayed dialog select the language of the original article and the article name, and the language you would like to translate to. Also add the title of the new article (or the original title will be inserted) and click on to begin. Your language preferences will be remembered for the next time.
  4. You will see a screen consisting of three columns. The left column contains the text of the source language and the middle column is for the translated text. Using the right column you can perform several actions such as insert source text, remove the inserted text source text, add or remove links etc.
  5. After you translate the article, you can publish it directly as a new page on the English Wikipedia by using the publish button that appears. In case the article gets created by another user while you were translating, you will see an option to save the newly published translation under your user namespace.

Users do not need to translate the entire article in one session. You can save the translation as a draft within Content Translation and publish it on-wiki when you are satisfied with the results.

Machine translation

Content Translation supports machine translation for a limited set of languages through Apertium, an open-source system. Machine translation is always available for use within Content Translation for any language supported through Apertium and for which we do not encounter any technical blockers. However, it is the user who makes the final choice whether they would like to use machine translation; it is a configurable option that users can choose to deactivate. With an ongoing survey, we are gathering feedback about the quality of machine translations so that we can work with Apertium to improve the service.

Content Translation is a quickly evolving tool, so we attempt to use all of the feedback we receive to improve the everyday experience of our users. Please send us your suggestions, comments and complaints on the Content Translation talk page or through Phabricator. At any time, the number of published pages and other details can be seen on Special:CXStats, the Content Translation stats page—see, for example, fr:Special:CXStats. This page is visible to all users of the wiki.

With the activation of translations into the English Wikipedia, there is a chance that there may be problems that we are not yet aware of. We will be monitoring the tool for these errors, but please do let us know on the Content Translation talk page or through Phabricator if you spot any problems.

Look for Content Translation in July.

For more information, please see the following pages:



The Wikimedia Foundation's Language Engineering team is "responsible for providing standards based internationalization and localization tools support for Wikimedia sites on the Web and mobile platforms." It is composed of six team members working on five different projects, including Content Translation.
The views expressed in this op-ed are those of the Language Engineering team alone; responses and critical commentary are invited in the comments. Editors wishing to submit their own op-ed should use our opinion desk.
+ Add a comment

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.

Excellent initiative. Great to see this going live in so many languages. Doc James (talk · contribs · email) 22:51, 25 June 2015 (UTC)[reply]

Waitaminute: the English-language Wikipedia is already the most bloated in the entire Wikimedia family, so you're giving us a tool to make it even larger? I thought this was going to be a tool for translating out of English into languages where articles are more needed! --Orange Mike | Talk 23:45, 25 June 2015 (UTC)[reply]
Many users can only write fluently in English anyway, why exclude them? The article count gives only an impression of better coverage, the English Wikipedia has a lot of stuff (much of it possibly of questionable interest, "bloated" as you say) but it's lacking on many important topics where other Wikipedias have better coverage; Mix'n'match is one place where you can find a list of todos. --Nemo 06:15, 26 June 2015 (UTC)[reply]
"Bloated"?! So, how many articles is supposed to be too many? Curly Turkey ¡gobble! 08:13, 26 June 2015 (UTC)[reply]
Maybe some greater coverage of Africa or Asia on En Wikipedia would be useful? Doc James (talk · contribs · email) 12:17, 26 June 2015 (UTC)[reply]
It is indeed incredibly bloated as far as pop culture and remote American villages with a population of 3 go, on the other hand there are crucial subjects regarding history and culture that are barely mentioned.--Catlemur (talk) 13:57, 29 June 2015 (UTC)[reply]
I read a while ago (but I cannot remember the source - help me if you can) that something close to 50% of the articles in the German Wikipedia don't exist in the English Wikipedia. That is, there is surprisingly low overlap of coverage even among the "big" languages. Also, as a side point to Orangemike's concern - translating an article from a small wikipedia into English (or even Simple English WP) is an excellent way to encourage that article to then be translated back "out" to other language wikis. Wittylama 12:50, 1 July 2015 (UTC)[reply]
You can follow all the articles published in en.wp with this special search on recent changes: https://en.wikipedia.org/w/index.php?title=Special:RecentChanges&tagfilter=contenttranslation It appears I missed out on my article Piazza Santo Stefano being the 1st mainspace article translation by a matter of seconds! Wittylama 17:28, 7 July 2015 (UTC)[reply]





       

The Signpost · written by many · served by Sinepost V0.9 · 🄯 CC-BY-SA 4.0