A number of rule-based visual languages have been developed over the past several years, including BITPICT [7], ChemTrains [4], Cocoa (formerly known as KidSim) [19], Cartoonist [12], and Agent Builder [17]. A visual representation is very natural for a rule-based language, for several reasons. First, individual rules map well to individual windows. Second, the independence of rules in a rule-based system [11] reduces the impact of the screen area required by a visual language, since only one or at most a few rules need to be displayed at a time. A third reason, in the case of the languages cited, is that rules identify local patterns within an array, and modify a small neighborhood.
One particular language that deserves special mention here is
LEGOsheets [8].
Like Altaira, LEGOsheets is a rule-based
visual language for reactive robot control. However, it ties rules to
the robot's individual effectors (rather than integrating all effector
changes for a situation in a single rule), and defines the rules
themselves textually. Also, implementing a hierarchy of rules to
provide a subsumption architecture in LEGOsheets would require careful
ordering of the conditions in the rule for each effector, in effect
writing a series of rules textually. In contrast, Altaira's rules
have a clear hierarchy and contain no textual component.