7.7 CONCLUSION

Current computer design environments provide opportunities to treat a product as a symbol and use this symbol for analysis long before physical prototypes are made. The product as a symbol is a center for information evaluation and enhancement. The DFM analysis of a CAD design for manufacturability is an example of how an artifact-centered approach can transcend engineering disciplines and create a design that allows designers and manufacturers to be cognizant of known problems long before the first physical prototype is produced.

The concept of DFM as an information processor can be extended past the design stage into the physical manufacture of a part. Information has long been treated as a by-product of the design and manufacturing process. The importance of such information has always been recognized. But inhibitions occur because of the traditional paradigm of placing the physical product, as it presently stands in the manufacturing process, bereft of history or future, as beholden to the production process. There is a limited, temporal view of how that product, or product line, has fared because concentration focuses only on the particular state of the product.

The product itself once was all that would be used to assess its condition and the processes that created it. It stood as the static representation of all the design and manufacturing energies that went into its creation. This paradigm has been carried to the point where a given object often is reworked until it is fixed, with little regard to its history and without providing a structure to accurately capture and analyze the additional detailed costs incurred to correct it.

With expert systems such as PIMES and relational databases which can stretch across all aspects of a corporation, a paradigm based on comprehensive information management will increasingly be of foremost importance within manufacturing organizations. Information can be obtained across traditional boundaries of organization and time. And expertise from multiple engineering domains can simultaneously be accessed without the attendant engineers.

However, information and expert systems which automate the design and manufacturing process should be recognized as tools for the engineer and not his replacement. Information design and management policies must acknowledge the presence of people as essential within the manufacturing process.

The use of symbolic processing, by its nature, can induce the engineer to be more amenable to expert systems. The symbols, inference processes, and results can be directly understood by the engineer as if they were his own. And the engineer may easily update the symbol systems himself to reflect changes in product or process. In essence, the engineer can intellectually grasp and reason with the same set of symbols used by his computer. More importantly, the computer can reason with the same set of symbols and heuristics used by the engineer.

Computers provide the capability to concurrently and effectively address the concerns of multiple engineering domains. The importance of computational symbols in automated engineering and engineering management should not be undervalued. Large-scale management of information requires research into extensive symbolic computation. At the same time, the presence of man in the manufacturing process must not be overlooked. The computational symbols used may be as important to man as well as machine. Although this paper focuses on the symbols such as features that are essential to the reasoning of the molding expert, it generally illustrates how symbol processing systems such as DFM and PIMES allow a more natural communication of knowledge between man and machine.