To clarify some thoughts and clear up some confusion, here is some definitions...
Definition: Meta Data
Information about information. Basically its documentation about information used or contained by various systems. Example: Data Dictionaries & Data Model Diagrams.
This is on excellent thing to do at the beginning of a project, no matter how small a project it is. See my thoughts on data modeling in previous postings.
Pipeline (ETL) processes can also use Meta Data to inform it of what is coming into the pipeline and how to process it.
Definition: Meta Data Model (Meta Model)
A flexable data model schema that can store the definition of the data it contains and may include relationship rules and constraints. This could be as simple name-value pair storage design to an elaberate rule based model such as the system tables (example: sysobjects) that SQL Server uses to store the definition of user defined tables and their constraints.
I have several comments about Meta Data Models in previous postings. And its to this particular definition of which I speak of.
Definition: Data Repository
A storage facility that stores data about data. Used by major corporations to track its many data sources. Its stores the data dictionaries, data models, and data flow. Repositories usually use a Meta Data Model schema design for its storage to organize the data.
Large corporate data warehouses may find this to be useful in keeping all those data dictionaries, diagrams, etc... organized. Usually not necessary for smaller businesses.