Model Management is a powerful approach to generic metadata management that manipulates  models and mappings between models using high-level operators. It aims at simplifying the development of metadata-intensive applications, such as data integration, software engineering, website management, or network modeling applications. Such applications manipulate a variety of 

• models (database schemas, XML schemas, UML / ER)diagrams, ontologies, etc.) and
• mappings between models (SQL view definitions, XSLT transformations,   XML-to-relational shredding specifications, ER-to-SQL DDL mappings, etc.).  

Model Management is a powerful approach to generic metadata management not limited to a specific language or application domain.  Models and mappings are manipulated using high-level algebraic operators, such as Match, Merge, or Compose. These operators are applied to models and mappings as a whole rather than to their individual building blocks. This approach, which was proposed by Phil Bernstein et al., promises to make the programming of metadata-intensive applications substantially easier.

Some of our key contributions are:

• Study of scenarios related to data warehousing to demonstrate the usefulness of model management (ER 2000)
• Development of the first prototype implementation of a complete programming environment for model- management, called Rondo, and its use to solve several realistic metadata problems (SIGMOD 2003). An executable demo of Rondo is available for download. 

Related Panel:

Bernstein, P.A., Is Generic Data Management Feasible? Panel discussion, Proc. VLDB 2000, pp. 660-662