A system and a method are described for reading data from a document and a method of detecting data. To read data which are present in volatile (i.e. time-variable) and weakly structured documents (e.g. HTML pages), a system with a computer having indicator and input means for operation by a user is proposed. The computer can access at least one document. The computer runs a program by which at least one first input by the user is processed with which a data range of the document to be read is fixed. The program automatically creates an extract instruction for reading the fixed data range. The extract instruction can, for example, be stored. When executing the extract instruction, the fixed data range of the document is read. This can be iterated with time intervals.
