Saturday, May 3, 2008

The Turjuman project

About 2 years ago, one of my friends was discussing the problem with how IT related books are translated. He stressed so much on how weak the translations are to the extent that at times the meaning was totally messed up. The real problem lies in the absence of a standard for translating books to Arabic. Every writer, translates based on his own understanding to the words with no referral to any kind of reference for this.

Looking around the Internet, the people at arabeyes have done a good job in creating a wiki-based dictionary. Although, I value their effort in maintaining that wiki, but there seems to be a flaw with it and to be honest, this has been the problem with most English-Arabic dictionaries around.

Usually in Arabic, a word can be used in several contexts with different meanings. This surely implies the need for a tree structure of the possible contexts ( of course the more in-depth , the less ambiguety ) to classify the words. Unfortunately, none of the efforts I have passed by solved this problem, they usually had a typical word-word translation. They had to attempt to get the most general translation that can fit most contexts, but then, that isn't applicable in all categories.

Several months later from my friends discussion, the idea came up again of starting this project and researching it, we came up with some ideas and and attempted to prototype some of them using drupal. We then documented some things on a public wiki and codenamed it, Turjuman.

Besides having a complete taxonomy for the translation, we touched the issue of the translation process. Who has the right to change the translation of any word ? Experts ? Anyone ? We had some thoughts about this. For example, maintaining a list votes on the user-submitted translations and the top one would be the *current* translation. This may be susceptible to many flaws, such as the equality of votes between expert translators and amateurs. This may call for a committee where their votes are factored up above others. Again, maybe that may lead to a tight dictatorship by the committee, opinions vary here.

Currently, we will try to come up with a simple prototype that would demonstrate the missing features mentioned above.

So what are your suggestions ? Feel free to add your comments.

No comments:

Post a Comment

What's your opinion on what I wrote here ?