Everything in LaTeX2Web revolves around the internal model of the document.

The model is object-oriented. Each class corresponds roughly to a specific component of a document: paragraph, list, table, etc... , and is also linked to a document description language.

Each class implements at least one of the following methods:

• import: this parses the document description code, and produces an object for the internal model as a result.
• export: this converts a model object into code for the corresponding language

Three document description languages are currently supported:

• Markdown
• LaTeX
• HTML

Document objects are not created from HTML, because HTML is such a general purpose language that we cannot fit general HTML code into one of the existing classes of the document model.

All classes do not export to markdown code because the language does not have all the features of the document model.

LaTeX code and the document model can be converted in both directions. In LaTeX2Web, there are specific extensions to the LaTeX language which correspond to specific web widgets.

This is a very concise description of how LaTeX2Web works, but it is helpful for understanding some of the advanced features of LaTeX2Web.