6.2 SIBTRAN (continued)

... a formal visual language was devised. This language, called SIBTRAN, formalizes a limited subset of the visual imagery system developed by Sibbet. SIBTRAN is a fixed set of graphic devices to organize textual sentence fragments under a system which provides a layer of additional semantic information (beyond the meaning of the words in the fragments). A SIBTRAN expression consists of one or more of the graphic elements (hollow arrows, bullets or straight arrows) placed in specified spatial relationships with either pieces of text (sentence fragments of typically six words or less) or other SIBTRAN expressions. The visual grammar for SIBTRAN expressions is shown in Figure 8. The parse and interpretation (text translation derived from meanings used by Sibbet in his work) for a standard SIBTRAN expression were presented back in Figures 2 and 3. This leaves us free to show an extension to SIBTRAN in Figure 8: expressions from two other formal visual languages, Bar charts and Feature Structures, are defined in the grammar as proper SIBTRAN expressions. Here Visual Grammar Notation allows us to express heterogeneous embedding, using SIBTRAN as a meta-level schema. Figure 8 shows the parse and interpretation for a mixed language SIBTRAN expression with a Feature Structure phrase to the right of the hollow arrow.

Action taken based on the parse: The SIBTRAN-assistant is a helpful interactive software module designed to facilitate graphic communication. The assistant uses the grammar in recognition and parsing. Its job is first to recognize when an arrangement of graphic objects is a SIBTRAN expression, and then to issue special prompts depending on the identity of the expression. The functioning of the SIBTRAN-assistant and its bearing on the conversational graphics problem (Section 7) is discussed in [Lakin86a,86b].