Meta data management is to data what data is to real life.
Data reflects real life transactions, events and relationships. Meta data reflects all of these in a purely data context.
The value of meta data management is in providing rich context for data so creating a taxonomy which is understood across the entire organisation. This, in turn, should supply the context for analytics, research and other search based functions.
Meta data is simply a card catalogue (like those used in a library) implemented in a managed data environment. If is therefore a very wide ranging term that can include may component areas, including business architecture, rules and definitions, data governance, integration and quality, document content, information technology and all sorts of data and process models.
Meta data is divided in four major types; business, technical and operational, process and data stewardship. It has a remit to cover structured and unstructured data, and as such is a bridge between the two. Meta data can have any electronic source whether that’s an entity stored in a database table or a scanned document on a users hard drive.
The architecture of meta data should be part of the data architecture function and the maintenance embedded within the data governance practice. These should define and assure how meta data is collected, the business and technical model of how it should be stored, the metrics to measure its quality and the standards it adheres to. The latter is especially important in the wider HEDIIP context as meta and reference data will be vital to a shared sector data language.
It can be confusing to understand where reference, master and meta data ‘fit’. This model (reproduced from Infogrid) is a simple explanation of all the terms you are likely to hear:
Complete machine readable specification for all data elements, properties and relationship types. Formal business rules.
Complete data model with inheritance hierarchy. Formal tree structure with formally defined concepts and properties.
Representations and relationships between data elements, including abstractions. Start of a 'common view' of data for the whole organisation.
Representations of vocabulary/glossary expressed within business process and - possible IT/IS systems. Data types, validation, formal data definitions.
Adding naming conventions and context to the glossary. Should include basic governance processes around change of these.
A simple list of terms and definitions. Complete list of terminology and acronyms.