You can program against an information model that is installed in a repository database. Programming against an information model adds, updates, removes, and retrieves data from a repository database.
Typically, the data that you add and otherwise manipulate is design data about a tool or application that you create. Furthermore, the data that you can add and manipulate is defined by the information model. You can think of the information model as a template to which the data you add must conform. For example, to create an application that manipulates a schema, tables, and columns, you need an information model that defines what a schema is, what a table is, and what a column is.
As a programmer who is coding such an application, you populate the schema, table, and column types with meta data instances to be used by the tools and applications you create. The following example provides a simplistic look at how you can program elements of a database application using the Open Information Model (OIM).
| What OIM defines | What you create |
|---|---|
| Schema | A schema for a Microsoft® SQL Server™ database, a Microsoft Jet database, or a new version of each database. In this case, four instances of Schema are stored in your information model. |
| Tables | Tables for Customers, Orders, and Products. For example, you can vary the table definitions based on the schema types, or you can reuse the tables for each schema. Creating separate tables for each schema results in twelve instances of tables in your information model. |
| Columns | Columns for Customer, Order, and Product tables. Assuming no reuse strategy, you can have a separate column instance for each table and for each schema. |
Notice that the instance data you store is all about definitions. Instead of storing "Joe Smith" customer name, you store data about the CustomerName column.
Meta data is, by definition, unbiased. The following suggestions describe different ways to reuse meta data.