The growing needs to analyse and interpret large amounts of complex information has generalised the use of information about information, often called metainformation (or metadata). Metadata approaches and standards have proliferated in fields as diverse as medicine, meteorology, geography, cultural heritage or education, among others. These approaches are supposed to assist us in documenting our information by recording who has documented what, when and how, among other concerns, making the tasks of interpreting the data much easier. However, metadata approaches often suffer from a number of issues. To start with, there are too many, and users are often daunted by the task to choose among them. Secondly, metadata approaches seem to re-invent the wheel by assuming that metadata is essentially different to data (or metainformation to information) and for this reason needs a new and different set of languages and tools. Finally, many metadata approaches mix together conceptual concerns and implementation issues, thus violating well-known engineering principles of modularity and layering. This paper presents a review of existing metadata approaches from a conceptual modelling perspective, identifies the major issues with them, and proposes a new approach based on the ConML conceptual modelling language. This new approach starts from the basis that metainformation is a particular kind of information and, as a consequence, everything that we know about information can also be applied to metainformation.
Metadata. Metainformation. Conceptual modelling. ConML.