ABSTRACT
Representing and reasoning with engineering design standards, regulations and codes is a highly active research area. Various methods and environments have been proposed and developed in the last 30 years. The existing proposals for standards processing include: the SASE Model (Fenves et al., 1987); knowledge-based (k-b) models (Lopez and Elam, 1989; Topping and Kumar, 1989; Garrett et al., 1995); logic-based models for standard representation (Jain et al., 1989; Rasdorf and Lakmazaheri, 1990); objectoriented models (Garrett and Hakim, 1992; Badrah et al., 1998); description logic models (Hakim and Garrett, 1993); hybrid models (Yabuki and Law, 1993; Neilson and Kumar, 1998); context-oriented models (Kiliccote et al., 1994); and fuzzy logic-based models (Schnellenbach-Held and Albert, 2000). A more detailed description and history of most of the earlier methods and environments can be found in (Fenves et al., 1995). Most of these methods and environments employ a single representation and reasoning method with a single locus of control, focus of attention and knowledge base. Each of these representation and reasoning methods has advantages and disadvantages compared to the other proposed methods. There are certainly other representation and reasoning methods that have not yet been used in standards processing. However, as identified by Kiliccote (1997), there is no single representation and reasoning method that is general, powerful and complete enough to represent and reason with complex design standards, such as the BOCA National Building Code (BOCA, 1993). Even if a unifying representation and reasoning method were developed, it would be difficult to use this single representation to represent and reason with all parts of a design standard. Also, as identified by various researchers (Guha and Lenat, 1990; Johnson and Mead, 1991; Sowa, 1991), the development of general representation and reasoning methods is not simply a matter of scaling up from simple ones.
