Sharnyo, what if Transwhiz provided an inaccurate "gist", or a misleading one? What are the consequences of getting a wrong interpretation? Will it be serious? This is a risk that you have to bear or consider if you decide to use it.
I have a cousin who lives in China's factory belt, down south in Guangdong. He speaks Mandarin quite well, having lived in Taiwan for almost 10 years. He has never tried to read Chinese and his work consists of building relationships with traders in UK or Europe so most of his communication is in English or Russian. Now a factory in China cannot possibly run without anyone in "the management" being able to read Chinese! Because invariably you are bound to have the utility bills like: gas, electricity etc.. etc.., provincial govt communications, the occasional restaurant menus and bills etc.. all written in Chinese. Since it is too expensive to hire someone who is say 100% bi-ligual, what my cousin does is hire a (pretty) secretary, to help out. First you categorise all the different documents that come to the factory and only mark "Red" for any unusual or infrequently seen items. Such Red items are discussed once a month at the "board meeting" and depending upon the "urgency" the board meeting can be scheduled earlier or later.
Now I would say, the secretary does a fine job of "translating" because she "learns" as she goes, noting eg. which documents are important, which are urgent, which items the boss must know etc. Even though her salary is about the cost of a copy of TransWhiz per month (or maybe even less!), after a few months, she can even help out with other issues not related to document handling. My cousin of course no longer has to worry about "exceptions handling" and focus 100% on his customers abroad.
Even though some people say Chinese is a language which only uses about 3000 characters to express near 99% of all written communication, the reality is that most Chinese words are 2 characters or bi-syllabic. What this means it that if you want to write a computer program that captures all the possible meanings that these 3000 characters "in combination" of one, two or three or more characters represent is truly an enormous task. We are also talking about mapping symbols which have not only a "one to many" mapping relationship to the target language but a "many to many" mapping relationship!
Gato, thanks for the articles on Google /Mit review of translation software etc... I found them quite interesting. I am sure it is only a matter of time, with so much computing power at out disposable, mankind will eventually be able to tackle this problem. In the future we will be able to shrink the whole program down into a C3P-O! Or just call the robot Data ala Star Trek.