Schemas

A schema, as one would expect to find in a database management system (DBMS), is a data model of the information stored in the database.

What typically distinguishes schemas from more general data models is that schemas may provide storage information that is irrelevant to the data model itself.

Examples include index definition and where database files reside on the file system.

Common types of schemas used in software

• Relational schemas
  These are much like E-R models, except that they can be specific about storage structures.

  Example: SQL Data Manipulation Language
  It has structure, but is not convenient for software to parse the argument's structure.

• Object-oriented database schemas
  These are much like object-oriented data models.

  Additional implementation details may include the source code of the methods associated with classes or instances of objects.

  Example: CORBA
  In CORBA, bindings are generated between details specified in a generic description of objects' interfaces (IDL) and the actual program code that implements those objects. That binding code, as well as the bound-to code itself, could reasonably be considered part of the schema.

• Semistructured schemas
  These are much like hierarchical data models. The additional details provided might include a canonical name for the schema (to be referred to by individual database instances), or what character set encoding is used in a database conforming to the schema.

  Example: XML DTDs

• Unstructured
  Many documents have structure that's easily recognizable to humans, but it very unclear to software. From a software system's perspective, these documents are unstructured.

  Examples

  • Plato's Republic It has structure, but is not convenient for software to parse the argument's structure.
  • A bitmap image of a molecule's shape The image contains structure, but it's quite difficult for an application to appreciate that fact.