Abstract: A method of processing semistructured data, in particular semistructured textual data, to output data which is in accordance with a predetermined structure, wherein said semistructured data is structured into one or more elements according to a given syntax, the actual content of the syntax elements being variable and being called a token, said method comprising: extracting by means of an extractor (“parser”) from said semistructured data one or more tokens, said parser being capable of returning at least one token in response to a respective specific command identifying the requested token by a token identifier, wherein said method further comprises: providing a sequence of commands and an associated data structure definition, both together being called a loader, said loader comprising the commands necessary to cause said parser to return the one or more tokens to be extracted; causing by said sequence of commands of said loader said parser to extract said one or more tokens from said semistructu