Conceptual Modeling of Events for Active Information Systems - Distributed Artificial Intelligence, Agent Technology, and Collaborative Applications - page 262

introduction
Organizations and the information systems that support them are artificial and intentionally designed
artifacts. Policies and procedures created by an organization determine how specifically defined and
identified events affect the organization. Active information systems are designed to participate in the
operation and management of organizations by implementing such policies and procedures. Events
play a crucial role in such organizational processes. They are defined and identified for the purpose of
initiating organizational processes among interacting participants.
The event commonly described as “the placement of a purchase order,” for example, is an intentional
agreement between a customer and a vendor. It is artificially identified and used by each participant to
initiate organizational processes. For instance, the vendor may create a conceptual object referred to
as a sales order—which is identified by a sales order number—and described by the particulars of the
agreement (e.g., payment terms, promised delivery date, FOB point, quantities and prices of products
sold, sales tax rate, freight charges). The sales order may also initiate production, shipping, and billing
processes. The customer, on the other hand, may create a conceptual object referred to as a purchase
order, identified by a purchase order number and described by the particulars of the agreement. The
purchase order may also initiate production and sales processes that depend on the receipt of the products
on that purchase order. It may also initiate processes to reserve cash required to pay for the purchase.
The ontological definition of an “event” as a state-transition (Bunge 1977) has been widely used in
conceptual modeling research (Wand and Weber 1990; Shanks, Tansley, and Weber 2003). This defini-
tion has resulted in the premise that an information system is fundamentally a state-tracking mechanism
(Wand and Weber 1990). It proscribes the representation of events as entities (Wand, Story, and Weber
1999). A conceptual model based on such a premise can appropriately represent a passive information
system (March and Allen 2007a), but it is inadequate in representing an active information system.
Effective analysis and design of active information systems requires a more substantive ontological
definition of an “event” as an identified causal occurrence (Geerts and McCarthy 2002; Davidson 1980;
Casati and Varzi 1996; March and Allen 2007b). Such a definition results in the premise that an infor-
mation system is fundamentally an event-processing mechanism (Allen and March 2003; March and
Allen 2007a). It requires the representation of events as entities for those events in which the informa-
tion system actively participates. The descriptions of such entities include (a) the organizational rules
governing the event processing and (b) the data that describe the event. We contend that it is appropri-
ate to represent events as entities at the conceptual level and argue that doing so is fundamental to the
conceptual modeling of information systems that actively participate in organizational work systems
(Alter 2003; Alter 2006).
Objects are described by functions that map to values. Such functions may represent properties that
objects naturally possess, or they may represent attributes that are artificially ascribed to them. The set
of values for the functions of an object at a point in time defines its state at that point in time. A person,
for example, possesses the properties height and weight and is ascribed the artificial attribute's name
and Social Security number. The mapping functions for a particular person may map that person to the
vector (height: 6 feet; weight: 230 lbs; name: Fredrick J. Smith; Social Security number: 001-01-0001)
on March 15, 2006. Note that the values of height and weight are each observable: they can be measured
at a point in time and they change according to natural laws (possibly influenced by behavior but ulti-
mately outside of human control). The values of name and Social Security number are not observable:
they cannot be measured per se and they are ascribed to the person and can be changed according to