A method of transforming and canonicalizing semantically structured data
includes obtaining data from a network of computers, applying text
patterns to the obtained data and placing the data in a first data file,
providing a second data file containing the obtained data in a uniform
format, and generating interface specific sentences from the data in the
second data file.