![[ ^ ]](active-up.gif)
![[ Top ]](top.gif)
![[ Home ]](home.gif)
Computers, by definition, translate one language into another. As we cannot speak the 0s and 1s that computers manipulate, they cannot directly manipulate languages which we can understand. To aid in the communication with computers, man formal languages have been developed (sometimes with a formal definition in Backus-Naur Format). These languages are often called "computer" languages and are part mathematical in nature and part linguistic in natures. They are designed to give an English-like interface to the computer. However, even statements such as:
is not always as clear as we would like. Further, as languages grow more complex, both the computer and the programmer must know more to translate from the desired results (the English project proposal, for example) into the computer's binary code.
Ideally, a computer would understand the spoken or written word, and if we were to tell it "give me the sales summary for 1989, 1990, and 1991" it would automatically gather the information which we need. Further, we would not even need to use English or even a specific syntactical structure to be understood. Granted, this goal is a long way from realization, but computers are getting better at understanding one word commands and responding to us in clear speech.
What is needed is to not formalize a natural language (which would, by definition, freeze it), but to make it so that a formal-mathematical machine can understand a natural language. The mystery behind this might be tied closely to the mystery of language acquisition; if we could understand that process better, we might be able to emulate it in Intelligent Agents on the computer to permit us to interact better with the information that is there.
. Get your own Free Home Page.
![[ Mail ]](mail.gif)
fantome@usa.net