Fig. 1. PDF document with its internal components. The PDF file layout consists of indexed pages, objects, and streams. The streams hold lengthy data for text and images.

Fig. 2. PDF document with a knowledge base in addition to XMP metadata. RDF/RDFS statements and OWL individuals in this knowledge base can refer to PDF annotations in documents.

Fig. 3. Annotation of PDF documents. It is possible to create notes and highlight text in the document.

Fig. 4. Overview of the PDFTab user interface. (a) In Protégé, users can select which tab to interact with. (b) PDFTab shows a list of documents associated with the ontology. From this list, users can select documents for viewing and annotation. (c) Acrobat runs inside PDFTab and displays the active document in a separate area.

Fig. 5. Document annotation. (a) The annotation tool. (b) Text annotation (highlighted). (c) The Protégé form for the OWL annotation individual.

Fig. 6. Annotation individuals. These OWL individuals link the PDF annotations to the ontology.

Fig. 7. Ontologies supporting document and domain modeling. (a) A sample annotation with corresponding links to the individuals. (b) The range of the annotation, document, and domain ontologies. (c) The boundary between Acrobat and Protégé systems.

Fig. 8. Alternative development processes for semantic documents. (a) Development of semantic documents from preexisting documents. (b) Development or reuse of domain ontologies followed by document authoring and annotation. (c) Generation of documents with annotations from domain ontologies.