Urdu Machine Translation Support

By Akbar

First, I must acknowledge that there are some very good Urdu Machine Translators are available on the web. They are not perfect and in a way it’s almost difficult, if not impossible, to build a perfect machine translator of any language, but still very good work has been done and it’s improving. Here are few I want to give credit for what they are doing:
http://www.paktranslations.com/
Apni Urdu

Now the one main limitation of these is that you must go to their site to do the translation. I have been working to add the Translations support in my Dynamic Language Tools bookmarklet, but this tool uses the Google Language API for the translation and for some reason Google haven’t yet added support for the Urdu Translation. Why, I’m not sure.  AWahid have some very good statistics on the Urdu Language popularity, but still not translation support is yet available from Google.

Anyway, as there is no API available for the Urdu Translation (as long as I know), I decide to apply a workaround for this. The google API does support Any Language to Hindi translation. As I explained in my previous blog that Hindi does share of most of the grammer and language rules with Urdu.  Now as I already have developed a good Hindi to Urdu Transliterator, I decided to give it a shot. What I have done is that when a user performs the Any Language To Urdu Translation, it internally translate that to the Hindi first, and then automatically transliterate that Hindi to Urdu.

The results are good, but there is still one issue. As Hindi uses most of the vocabulary from the Sanskrit, while Urdu vocabulary is mostly from Persian, Arabic and English. Thus the text translated this way is not very readable to native Urdu speaker. I found and implemented a workaround for this too, and that’s if you double click on any translated word, this tool also brings and displays the corresponding words from the Arabic, Persian and Arabic for that selected word as shown below:

Urdu Translation Options

Now as you can see, the Good Morning was translated to “????????“. As Urdu native speaker, I haven’t heard this word before, and we in Urdu use the word”??? ????” for this, and this is what is visible in the word options (it’s a Persian word for Good morning).

Concluding, though the automatic translation is not great, and user may need to replace most of the Sanskrit words, but it’s just the first version. Hopefully, it will improve as I get some time and ideas.

Please do keep sending you feedback and comments.

Tags: ,