Documents & Views

  
Figure 1: LINCKS Table of Contents View

A document (or rather a document view) corresponds to the content of one HTML page or one window in xlincks. A typical instance of a table of contents from xlincks can be found in figure 1 and the corresponding page from WWW is in figure 4. The view is constructed from several objects. Each section title is contained in a separate object, thus the root object is a composite object. In LINCKS we construct a view by applying a general presentation descriptor (GPD) to a root object (often a composite object).

General Presentation Descriptor (GPD)

The general presentation descriptor contains no actual data but provides the framework for both retrieving and storing the different parts of an object. The information in the general presentation descriptor is divided in following parts [PL93]:

• The STRUCTURE part which defines the logical structure of the view and its decomposition into items or structures. It similar to the DTD in SGML [Gol90], but much simpler.

• The ACCESS part defines where in the database we find the information we need to construct the reference structure.

• The EXPAND part defines which GPD should be used when following a hypertext link (e.g. expanding an object).

• The EDIT MODE part defines restrictions on the terminal items - if reading, writing, expanding and linking from is allowed.

• The FORMAT specifies X11 format attributes used by xlincks - font, borders, margins etc

• The FORMATX specifies conditions and actions to generate textual output (HTML or [Lam86]) from the reference structure.

  
Figure: GPD Template for `TOC View'

The reference structure represents a view of a particular object as it exists at a particular time. It is built using a template and actual data objects, the content objects. The template itself is constructed from the structure and access parts in the GPD.

  
Figure: Reference Structure for `TOC View'

If we apply the template in figure 2 to a root object, we end up with the reference structure in figure 3. A reference structure is thus an instantiated template and has the form of an ordered tree with named branches. The view in figure 1 is the result of displaying the leaves of the corresponding reference structure (in figure 3).

  
Figure 4: WWW Table of Contents View

Composite Objects & Multiple Views

Composite objects are built using a general presentation descriptor, which provides a template for a composite class (e.g. the links have the part-of intuition), and contents objects. The contents objects are the root and the components of the composition. This way of handling compositions gives us a strong flexibility. It allows us to share contents objects between different compositions and allows different views of compositions sharing the same root. It also allows us to use different GPD with the same root object, giving us different compositions (or views of a composition) sharing the same root - see figure 4 and figure 5.

  
Figure 5: WWW Alternate Table of Contents View