LangPrism - a web content crowdsourcing translation service


    Hello, dear habrasociety!

    I want to introduce you to our LangPrism Internet translation project . Crowdsourcing has long and confidently entered into all spheres of human activity and the use of crowdsourcing for such an area as translation is not something particularly new, however we decided to look at it from a slightly different angle and want to present you our point of view in the form of our project . Perhaps many people are already familiar with our translations here on Habré ( translation , translation , another translation ), but, nevertheless, I decided to publish a separate post to tell about it in more detail.

    Under the cut you will also learn about the new Habré in Japanese , do not miss!


    Praeteritum


    About two years ago, Microsoft ran an ITBonus loyalty program among its community , in which it encouraged various activities related to Microsoft products. Among them was the translation of technical articles and blogs on Microsoft technologies. Then I actively participated in this program and even a couple of times I was in first place in the ranking. However, from the very beginning, I was very upset by the inability to share my translation. There was no other way but to send the html-file with the translation to the organizers by mail. Not to mention the fact that you also had to manually edit the translation using the developer tools in Chrome.
    Then the idea came up, and why not make such a service that would allow the translator to change its content directly in the context of the web page, and other users to view this already translated content. Having developed this idea a bit, it became clear that it is very similar to the Wikipedia concept.

    Big Idea


    Our big idea is to combine the idea of ​​Wikipedia and Google Translate. Literally, you visit a website in a foreign language, use LangPrism, and if you have the opportunity to write an encyclopedic article on Wikipedia, then LangPrism allows you to contribute by translating the web page into another language. At the same time, your translation instantly becomes available to other users, which allows them to read foreign web publications in their native language.



    This all works due to the LangPrism extension installed in the browser, which exists for all modern browsers: the native extension for Chrome and the extension in the form of a bookmarklet for other browsers. When the extension is activated, the following occurs:

    • The contents of the original page are sent to the server.
    • The server returns fragments of this page already translated by users.
    • Finally, you get the translation user interface directly on the source page and you can translate by right-clicking on sentences.

    Once a sentence is translated on some website, it remains in the translation memory system and can be obtained from there upon further requests. Thus, after some time, the system itself will play the role of a global translation memory, which can be used not only for translating web content, but also in other areas, for example, when translating technical texts, instructions, etc.

    Praesens


    Now about 1,400 people use the LangPrism extension for Chrome , the number of registered users is about 420. Anonymous users also make a significant contribution to the translation, translating up to 100,000 words per month. The community itself selects the sites for translation based on its preferences, the distribution of the amount of translated content by domain now looks as follows:



    In total, more than 400,000 words have been translated during the project’s existence, which is comparable in volume to two texts of the War and Peace or five solid 300 page books. This is pretty good considering that the writing cycle for one book is about 1 year. One of the most popular sites that our community translates to is TechCrunch.



    Today, the role of moderators in the system is performed by the users themselves and the simplest mechanism for combating spam. By analogy with Wikipedia, we made a post-moderated model for adding a translation, so that any of your translations immediately becomes available to other users. Now one of the directions of our work is the development of an intelligent technology to counteract the distortion of meaning during translation, which will allow us to more accurately control the quality of the crowdsourcing translation process.



    More recently, the section appeared on the service Translations, which includes the most popular foreign articles found on the Internet. It provides an opportunity for translators to translate relevant articles and allow other people to keep abreast of all the most interesting things that appear on the network. The section itself is designed like now popular aggregation services of interesting content. A sort of StumbleUpon in Russian. Thanks for the help in developing UX, Chipsadesign .



    The flip side of all the visible translation work on web pages is the LangPrism community portal. It is a kind of social network for translators, which allows them to solve interaction issues in joint translation, compile glossaries and agree on the style of translation. Each translator has a profile on the community portal. The portal provides a live feed that shows the latest translations added by users, and a peculiar rating of community users collected by the number of translated words.

    Harry Potter in Russian




    Now one of the most active areas of work of translators is the translation of the game based on books about Harry Potter Pottermore. Browser games turned out to be a great niche for using LangPrism, since there is no other way to “pull” localized content on them, if there is no official translation. Pottermore uses flash, but the main text is written in html, which makes it easy to translate. Translators used to do Pottermore translations, only posted translations in specialized forums, or in VKontakte groups. To use it, I had to open two browser windows and compare the source text with the translation in order to understand something. Now we are happy with the translators themselves, who have a convenient tool for translation, as well as ordinary players who call the incident nothing more than magic.

    Habr in Japanese




    Not so long ago, translators from Japan became interested in the service. Suddenly, they began to translate Habr into Japanese. You can try to use Yaponokhabr right now. If this trend continues, we have a chance soon to broadcast for Japanese habruzers in their native language! Taishiro , Yukihiro thanks for contributing!

    Futura


    The plans for the future are to make LangPrism a unique standard for exchanging multilingual information on the Internet, since now quite important and useful information has been created in different languages, which does not make sense to generate anew each time. It is enough to simply build informational, cultural bridges between different languages ​​and the total value accumulated throughout the world will be available to everyone. As Wikipedia once provided a tool that made it possible to make a cast of all the knowledge that humanity possesses, so we want to give the opportunity to bring this and many other information to every user on the Internet.

    I would like to invite all the habrayuzers to participate in this project, and if everyone who reads this article makes at least one translation, shares his idea or other help, we will take one more step closer to this truly great goal.
    Welcome to LangPrism .

    Also popular now: