Resumen
In the article, authors say that although Web information is meant to be consumed by human beings, applications that process it automatically are both conceivable and desirable. A useful abstraction for modeling these problems is to view the Web as a collection of, usually small and heterogeneous, databases, and to view programs that extract and process Web data automatically as database applications. WebOQL system, developed at the University of Toronto, is a tool for building the software layer through which such programs can view Web data in a uniform, convenient format and can effectively manipulate it using a declarative query language. In WebOQL, every data source is logically viewed as a graph of trees where every arc in a tree or graph is labeled with a record. The mapping between a data source and this logical view is performed by wrappers dynamically, as queries are executed. WebOQL queries can be embedded in Java programs like SQL queries are embedded in C programs