XML: A Managers Guide (2nd Edition) (Addison-Wesley Information Technology Series)

The convergence of information exchange problems reveals the importance of a general-purpose approach for structuring information exchanges. Such an approach would have to meet two requirements: (1) a party could indicate what each piece of information in a conversation means, and (2) two parties could agree on a limited vocabulary to use for the exchange. The first step in developing such an approach is to meet requirement (1) by adding metadata to documents.

Metadata is data about data. For instance, "author" is metadata about "William Shakespeare." A database schema comprises metadata about the data in the database. It gives all data in a column the same description ”such as "Customer Number" ”and indicates the datatype of data in that column, such as currency, integer, or string . A distributed object interface definition comprises metadata about the objects that implement that interface. It describes the behaviors an object implements and the arguments necessary to activate each behavior. Metadata is a formal way to describe what a piece of information means.

Metadata addresses requirement (1) for a general-purpose approach to information exchange. However, it does not address requirement (2). Metadata by itself would allow the home search site mentioned previously to describe a set of search results as useful data. But the user 's browser would not necessarily understand the organization of that description. If "Home Price" has the metadata "number" attached, what does that mean? What if the metadata were "numeric," "int," or "dollars"? Does the browser need a library of common terms for the same concept? Metadata alone certainly wouldn't solve the problem of business-to-business commerce. What if the metadata "price" attached to "100,000" meant U.S. dollars for one company and pounds sterling for another? Metadata makes it easier to locate interesting data and associate a piece of data with a description, but it still requires the explicit agreement of each party. The parties have to agree on the structure of the metadata.