[cnn-photo-caption image= http://i2.cdn.turner.com/cnn/2009/images/01/13/art.yamli.homepage.jpg
caption="Yamli.com allows users to search for any Arabic phrase they want."]
French president Nicolas Sarkozy recently called Arabic, “the language of the future, of science and of modernity,” according to the Brussels Journal. His comments coincide with the French government’s urging of Arabic language and civilization courses in French schools.
If you do speak Arabic or if you’re trying to learn like I am, then you know that it is an incredibly challenging language. Those challenges extend to the Web where the possibility of finding news and information in standard Arabic script is overwhelming and often incomplete.
Enter Yamli.com. It is an Arabic transliteration search engine that allows visitors to use their Latin keyboards to search for any Arabic phrase they want. Professors at major universities, including UC Berkeley and Stanford, use Yamli in their Arabic language classes.
“We created the site to solve the everyday problem of typing Arabic,” says Yamli co-founder, Habib Haddad. “A lot of the content on the Web today exists in two forms: standard Arabic script or messy Latin variations.”
Haddad says most Arab newspapers and professional blogs write in Arabic script. But he adds that a majority of user-generated content and comments on social networking sites are written in ad-hoc Latin script, which he dubs “Arabizi” or “Arabish” (a portmanteau word).
For example, here is the name of the famous Egyptian singer Umm Kulthoum written in Arabic:
أم كلثوم
Here are a few possible variations in Latin text:
Umm, Om, Oum, Im, Um, Em
Kulthoum, Koltoum, Kalthoum, Kolthom, Kalthoom, Kalthom, Kolthoum
Since there is no proper or exact phonetic match from Arabic to English (Latin text), it creates a widespread problem on the Web because multiple variations exist for a single word.
“It creates a vicious cycle in the Arabic search space,” explains Haddad. “It starts with the difficulty of typing Arabic, which leads to less people searching for it, and less money for Arabic publishers.”
Yamli offers a solution in that it allows users to spell the Arabic word the way they think it would be spelled in English. It automatically provides matches from that spelling and other relevant matches from alternate spellings.
Along with text, Yamli offers image search powered through Microsoft’s Live Search API and video search from YouTube. The search results are instantly presented in two columns: one side in English, the other in Arabic.
"We think we've solved some of the fundamental problems in Arabic search that have not been addressed by the big guys such as Google or Microsoft,” says Haddad. “We are really excited to already see its positive impact on users and believe it will set a new standard for Arabic search.”
Posted by: Valerie Streit -- CNN.com Producer
Filed under: computers • Internet • technology
