Browsing through the Global Voices ‘Most Read in 2012‘ articles, I noticed for the first time the effort that the site makes to accomodate multiple languages.
First, each article has a header saying ‘this post is also available in…’ Then, within the text, any quotes are presented in their original language, followed by a translation. This article on a Citizens’ Basic Income initiative in Switzerland is a good example.
Such efforts always introduce a design problem. The foreign language breaks the ‘flow’ of the text, much like a banner advert in the middle of an an article. On the Global Voices pages, the designers have made some slight attempt to slightly de-emphasise the original text in favour of the English translation. However, the long list of alternative languages at the tope of the article presents a barrier to actually reading the text, even when the type size is small.
Technical advances could help solve this problem. Automated tools like Google Translate allow you to put a rudimentary translation behind a single mouse click (you can read this very blog post in French, Arabic, Chinese or Hindi, for example).
However, it strikes me that some development of HTML standards to accomodate alternative languages would help. It is possible to embed extra data into any piece of text on a website. I ranted a bit last month about how quotes and links should include citation information within their code. The abbreviation tag (<abbr>) typically allows the writer to include more information for the reader, about a particluar acronym. An example of this feature in action: HTML.
It would be great if a ‘translation’ tag or an ‘original language’ tag were made available in HTML. Then, any given piece of text could be directly associated with any alternative language translations. This would allow web designers or app creators to display the translations in the way most appropriate to their content. It would also give readers the flexibility to show or hide the orginal language text.
Translation is a highly cultural and political act. How a piece of text is translated matters, and the nuance and tone imposed on the text by a sympathetic or antagonistic translator can have huge consequences. Technical innovations that allow the original and translated texts to be compared easily would mean more transparency, greater international and inter-cultural understanding, and perhaps even help students to learn foreign languages, and the art of translation. It may also help unify projects like Wikipedia, where there are almost three-hundred encyclopædias evolving in disparate directions.
Does anyone know if there has been any work or research done on this subject?
2 Replies to “Handling Translation on the Web”
Leave a Reply
You must be logged in to post a comment.
Really interesting post and an important topic. At Meedan we’ve worked on building hybrid translation workflows (MAT + human) and have experimented with displaying versions and original languages in different ways, perhaps the most interesting being the bold display of original language (which, to many users, will be meaningless – we liked the way it displays a back and forth between speakers of two different languages) in discussions such as https://news.meedan.net/index.php?page=events&post_id=316365
I like the idea of an HTML translation tag – lots of challenges and possibilites with regard to workflow, versioning, display elements etc. One project working on something similar, but with an added layer of complexity, is the UNL project at the Bibliotheca Alexandrina http://www.bibalex.org/unl/Frontend/home.aspx
As a professional translator, I take objection to your notion of sympathetic and antagonistic translators.The best of all professional translators abide by impartiality at all times, and the few that don’t generally end up getting caught out by a member of the vast majority who do.