||Pre-Quillpad(before 2006) transliteration technologies required one to type according to pre-defined rules. For eg: to get राष्ट्रभाषा, user might have to type rAShTrabhAShA. Such rigid case-sensitive typing rules, made using Indian languages on computers very unattractive for a layman.
A transliteration technology that allows users to type words as they
would casually do(like 'rashtrbhasha' instead of 'rAShTrbhAShA)becomes
the fundamental necessity.
||Try out something like 'sare jahan se achha hindostan hamara'
||सारे जहाँ से अच्छा हिंदोस्ताँ हमारा
Note: You don't have to write 'saare' or 'sAre'. Quillpad also naturally supports multiple ways of typing a same word.
For example in Hindi:
Will all give the same output.
|You are the best judge to decide who is better; so go ahead and copy-paste the test examples in the competing transliteration applications.
engine that can automatically learn to transliterate multiple languages
without the help of any human linguistics expert
intelligent transliteration system that works very well for one
language is not a good indicator of the technology behind it. Artificial
Intelligence(AI) based transliteration that works equally well for
multiple languages, without requiring any kind of language specific
customizations is a must.
||If any transliteration system supports only one or two languages, then it could be possible that the underlying technology may not adapt naturally to the other languages.
||The Quillpad Machine Learning module can learn a new language within a matter of few hours. The engine is not even specific to Indian languages. It can learn any language that can be transliterated using any other script, equally well. We don't need to change even a single line of code in our Quillpad Learning Engine to achieve that!
|Ability to transliterate English words that are typed with conventional spellings
||English words are so common in our vocabulary. Nobody speaks or writes in pure Hindi or Kannada
for eg. Also, when writing native English words, it's very likely that
we use their conventional spellings. A superior transliteration system
must automatically handle transliteration of native English words. Sub-standard transliteration tools support manually added limited set of frequently used English words or expect users to type English words phonetically!
||Try out 'News: Congress calls Anna Hazare unelected tyrant.
and also try out 'Jayalalitha looking for new equation with centre?'
We are sure you wouldn't prefer to type words like tyrant, equation, congress, etc... to be typed as 'tirant', 'ekuation', 'caangress' in order to transliterate them correctly!
|न्यूज़: कॉंग्रेस काल्स अन्ना हजारे अनेलेक्टेड टाइरांट.
जयललिता लुकिंग फॉर न्यू ईक्वेशन वित सेंटर?
Quillpad uses a superior mechanism to transform English words to Indian English way of writing in Indian languages, automatically. It is not based on manually added limited words set.
|Support for multiple languages
systems targeting countries such as India, having 10s of different
languages must scale automatically for all of them.
||Check how many languages are working effectively
||Best way to see how Quillpad works for multiple languages is to see what users are talking about it. Check out Quillpad Forums. Our developers know how to read or write only Hindi, Kannada and Telugu. But Quillpad is currently trained for 8 languages. We didn't need any linguistics expert to help us achieve this. Quillpad does it automatically, the Machine Learning way!
|Ability to work with other input interfaces like touch devices
||The 101-key keyboards are not the only devices for input; what about mobile devices with touch support? Imagine having your touch devices support SMSing in your mother tongue, as easily as they support typing English words.
if a transliteration system is capable of being ported to a touch device so that you could type your preferred language just like you would
type English words.
||Check out Quillpad Touch for a demo of how the Quill engine can work even with a touch device. That's the real proof of the generalizing capability of Quillpad technology. We were able to train Quillpad for the touch keypad, without having to modify the engine even by a byte!