Generation of Web Sites



(nur auf Englisch)

One adequate approach to produce a consistent Web site is to generate the site from data stored in one or more repositories. In addition to the data (content), descriptions of the site structure and its design are processed during the generation.

Using graph technology for the repositories seems to be a natural way, especially when it comes to represent instance data that is highly interconnected. Our TGraphs can be enhanced with schema contraints. Thus, the graph itself can be an instance of a conceptual model prescibing classes and their relationships.

This again is very interesting for Web site generation, for the conceptual domain world model is directly mapped to a graph schema. Instance data can be checked against this schema, and the site structure can be modelled as a graph as well. Defining the page contents involves graph querying.

This way, the Web site is defined by

  • a conceptual model (graph schema)
  • instance data (graph)
  • the site structure (graph)
  • views on the instance data (graph queries)

As an addition, the page design (presentation) is applied to the query results. Let the query results be XML documents, then the presentation can be one or more XSL-T stylesheets that transform the query results into HTML.


This image shows a typical Web site generation process.