When discussing standardization and secondary analysis in computer driven historical research, we should clearly distinguish between different approaches towards the usage of computers. The properties of four of them – statistical analysis, structured data bases, full text systems and annotation systems – are discussed and compared in some detail. Standards which want to convince users, that they should follow just one of these approaches will not succeed. What we need is a discussion of the communalities between the underlying information models and the identification of properties, for which clear conceptual models can be devised. Such clear conceptual models are a prerequisite for technical solutions, which ultimately can enable the exchange of data across the different approaches. Such conceptual models, therefore, are what we need as standards.