Monday, July 26, 2010

Some thoughts on a language neutral web

I have been playing around with Google translation and transliteration APIs which I found is very useful to overcome the language barriers in using the web. This especially have a lot more importance as web is still not accessible to a vast majority of the people in the developing world due to many reasons.Even if it is accessible language becomes a barrier for most of the people to utilize its benefits as only very little resources will be available in their local language.To make the idea of having the knowledge accessible to everyone everywhere at anytime possible it is important to solve this language problem.

Think about this idea of viewing every web page in your local language.You choose the language in which you want to view the page and the browser translates the content of the page to the language of your choice.I know some of you might already know that it is possible with google translation toolkit.But I was not fully happy with google translation as it is not meant for a common man who knows only his local language.First of all the interface is still in English and the user has to copy paste the URL into the box and do translate to view the translated page.So this is not aimed at a common man who only knows his local language(other than English). What I was thinking about is to have a comprehensive solution which will lead to the idea of a language neutral web. Let me put my thoughts into 5 basic points.

1) Have a browser interface in local language.
Solution :Yes it is already available in some form.For example Firefox have a localization option by which you can change the interface to the language you wish.This is good and satisfies most of my requirements .But still this is just a replacement of English with corresponding local language words and the interface and technical terms is not easy to understand for a man without fundamental computer knowledge.I wish to see such an interface existing.

2)The user should be able to search in his local language.That means a search interface in local language where he can type in his local language and do search.
Solution : Yes this is possible as Google has a option to choose your local language and search in that language.Of course there are many other search engines with same capability.

3) In addition to search results in local language it should also bring the pages in other languages that matches the word user typed in his language and this test results should also be translated and shown in his language .

Solution : This is not currently available with Google search(I assume the same with any other search).

4) When the user clicks on any search results(or may be direct access of the page by typing the url) that web page should be opened and shown translated in his language if it is in a different language.That means any web page the user access should be translated and shown in his language if it is in a different language.
Solution : There are some tools and plug-ins available for Firefox which provides an auto-translate feature.Examples are FoxLingo and Worldwide lexicon. They basically uses Google translation or any other translation services. I found Worldwide lexicon plug-in most suitable as per my requirement but it is bit slow as it takes some time to translate the page stage by stage.

5) The user should be able to type in any web page in his local language and if the original web page is in not in his language the text input should be converted to that language while submitting.
Solution : I know Google transliteration can be used to type in your local language in any web page.But there is no direct way for the automatic conversion of the text in input box to the language of the web page .Of course you can copy paste the text after conversion in an indirect way.

Challenges

1) Some of the contents of the web which are not in textual format cannot be translated.
Eg. Images,flash ,sound and video files.
Also documents embedded or down-loadable from the web should be converted separately for viewing in his local language.

2)Machine translation is not yet 100% correct even though there are efforts from multiple sources to improve this.Also it is available only for few languages right now and developing translation algorithms for all the languages is a challenge.For eg. Google translation services are available only for 57 languages right now.

Benefits

1)The huge benefit I see is in making available huge knowledge base like Wikipedia and other sites to every one irrespective of their language. This is important as local language version of Wikipedia content is very limited compared to that in English and other major languages.This helps very much in education also.
2)The web developers need not worry about localization of their websites and is easy to reach about to a wide range of people using different languages.
3) Most importantly users are not restricted to use the service of any website because the website is in a different language. He will be able to view the site in his language and also will be able to interact in his language.

So is it achievable?

Yes. To a large extend still keeping the challenges in mind(And of course considering only textual data at the moment). Also you can see that among the 5 points only points 3 and 5 is not available as of know(As per my understanding) and that is also achievable with the existing technologies.Only thing which we are lacking is a comprehensive solution integrating all the five points which we discussed and I am sure it is going to happen.I request readers to put your thoughts and also to point out anything which I might have missed out.

No comments:

Post a Comment