Wednesday, March 11, 2009

Multi-Lingual Systems

Sometimes, we browse content on the web and wish that there was an easy way for us to just switch between languages to make the experience more "wholesome". Most websites out there actually come with a language option varying between different dialects of the English language coupled with some other latin script languages. Of course, this offers users in those languages the opportunity to experience the website in the language of their choice but what about all the other languages that sit outside the scope of what those websites are offering? Do they simply pick up another language or do we assume that everyone reading these pages is a master of one of the choices on the page? The answer may lie in web technologies or future WWW technologies that do translations based on inputs received from users in the system. One such product that I have had the pleasure of working with is IBM's Websphere Translation Server. The server sits as an intermediary over Sametime (The internal messenger used by most IBMers) chats and converts the languages before they reach their destination. For eg. I type in English and the recipient sees it in German. However, this is certainly not enough considering the goal is still not accomplishing a total conversion task.

While I have pondered on the requirements that would make such a feat possible, the end result isn't really all that complicated when you consider the following points.
  • Most dictionaries are updated yearly (sometimes the frequency is much less)
  • All systems are moving to e-Storage
  • Most systems aren't allowing legacy storage techniques as the concept of ease-of-access gets disrupted
  • Languages continually change are modified by consortiums sitting in those countries
  • Web technologies like CSS3 and XML are making the user-defined experience more easier to deploy

The list can once again go on and on but lets focus on these points for now. The primary problem with language-related development is that there will always be significant differences in interpretation when one language is converted to another and this has to be considered when working with technical content as well as literary content. Here are some solutions that might be tried.
  • XML Style Sheets that use conversion methods in the framework for moving certain words and their combinations together
  • Developing world-wide standards for delivering webpages without hampering the process and creativity that goes into most websites
  • Encouraging more developers and designers to start using content development techniques that are feasible to easy conversions thereby touching more users in other languages
  • Enforcing a standard set of protocols in web-browsers for easy conversion that may eliminate the XML Style Sheets mentioned in the first point. The less work done on the website means less work by the developers
  • Using the systems managed by Linguistic teams globally to update the XML content and thereby offering users the most updated information
While the road is long and hard, I believe that most people will benefit from this development. Can you imagine a world where your content can be carried in your pocket on a mobile device and simply converted on call? How about working on a project in India and then deploying it in Japan without ever having to worry about the language requirements in the middle? Imagine a world where creative minds didn't need translation services and worked on simply taking content and pushing it out for everyone to read. The pros are many but unfortunately the amount of jobs lost and the continual man-hours spent in updating XML Content would surely outweigh this need right now. The next step I guess is to automate the updates ... One day.

Music Listened to while writing this blog - Joan Jett - Dirty Deeds (AC/DC Cover)
Mood - Creative