Wrong Trend vs. Wrong Focus CAD technology has been very successful in the last ten years. CAD tools for layout and logic design have been exceptionally successful to the point that they dominate system and chip design methodologies throughout the industry in the U.S. and abroad. This widespread methodology consists of manually refining product specifications through system and chip architecture until the design is finally captured on the logic level and simulated. Standard-cell methodology and tools were developed for easy rnapping of logic-level design into IC layout. Because of the huge investment in CAD tools, equipment and training, many people believe that this trend will continue by providing more sophisticated CAD tools for capture, simulation and synthesis of logic-level designs. Logic level, however, is not a natural level for system designers. For example, when we want to indicate that 32-bit values of two variables, a and 6, should be added and stored in the third variable, c, we simply write the expression c = a + 6. We do not write 32 Boolean expressions with up to 64 variables each to indicate this simple operation. It is very difficult to imagine having complex multi-chip systems described in terms of.l million or more Boolean equations. If we equate layout-level of abstraction (transistors, wires and contacts) with machine-level programming then logic-level (gates, fiip-fiops and finite-state machines) can be equated with assembly-level programming. We know that complex software systems consisting of 1 million or more lines of code are not written in assembly language. Similarly,a complexhardware system of 1 million or more gates should not be captured, simulated or tested on the logic level of