Human machine translation for research so far has been
for decades, but regrettably, can completely replace human
translation to machine translation system has not yet
appeared. As mentioned earlier, the current machine
translation systems, there are four, and rule-based, based on
statistics, examples and based on multi-engine. Wherein the
simple rule-based machine translation systems have proven
to be essentially unworkable, because human language is
profound, can by no means certain set of rules plus a number
of the dictionary can be exhaustive. The current popularity of
machine translation field or the statistics and examples of
multi-engine system together, in fact, in a multi-engine
system is the most important thing was the theory, "statistics
+ theory" whether multi-engine machine translation system
machine translation is correct now the direction is still
uncertain, but at least it has made good achievements, the
most typical example is Google's online translation system,
using this translation system has been able to get some on an
acceptable translation.