【正文】
teristics: ■ Design data is characterized by a large number of types, each with a small number of instances. Conventional databases are typically the opposite. For example, the DreamHome database consists of only a dozen or so relations, although relations such as PropertyForRent, Client, and Viewing may contain thousands of tuples. ■ Designs may be very large, perhaps consisting of millions of parts, often with many interdependent subsystem designs. ■ The design is not static but evolves through time. When a design change occurs, its implications must be propagated through all design representations. The dynamic nature of design may mean that some actions cannot be foreseen at the beginning. ■ Updates are fatreaching because of topological or functional relationships, tolerances, and so on. One change is likely to affect a large number of design objects. ■ Often, many design alternatives are being considered for each ponent, and the correct version for each part must be maintained. This involves some form of version control and 2 configuration management. ■ There may be hundreds of staff involved with the design, and they may work in parallel on multiple versions of a large design. Even so, the endproduct must be consistent and coordinated. This is sometimes referred to as cooperative engineering. Computeraided manufacturing ( CAM) A CAM database stores similar data to a CAD system, in addition to addition to data relating to discrete production (such as cars on an assembly line) and continuous production (such as chemical synthesis). For example, in chemical manufacturing there will be applications that monitor information about the state of the system, such as reactor Bessel temperatures, flow rates, and yields. There will also be applications that control various physical processes, such as opening valves, applying more heat to reactor vessels, and increasing the flow of cooling systems. These applications are often organized in a hierarchy, with a toplevel application monitoring the entire factory and lower=level applications monitoring individual manufacturing processes. These applications must respond in real time and be capable of adjusting processes to maintain optimum performance within tight tolerances. The applications use a bination of standard algorithms and custom rules to respond to different conditions. Operators may modify these rules occasionally to optimize performance based on plex historical data that the system has to maintain. In this example, the system has to maintain large volumes of data that is hierarchical in nature and maintain plex relationships between the data. It must also be able to rapidly navigate the data to review and respond to changes. Computeraided software engineering (CASE) A CASE database stores data relating to stages of the software development lifecycle: planning, requirements collection and analysis, design, implementation, testing, maintenance, and documentation. As with CAD, designs may be extremely large, and cooperative engineering is the norm. For example, software configuration management tools allow concurrent sharing of project design, code, and documentation. They also track the dependencies between these ponents and assist with change management. Project management tools facilitate the coordination of various project management activities, such as the scheduling of potentially highly plex interdependent tasks, cost estimation,