Several common themes emerged from panelists' opening statements. One, that engineering education had improved in many ways since the panelists themselves received one, was nearly unanimous. But it was equally agreed that the improvements have not kept up with the growing challenges faced by engineers and their managers.

One frequently made point was that neither an education of pure theory nor one of cookbook solutions would suffice. Charles Sodini of MIT suggested that education needed three components: design, simulation and experiment. While agreeing that engineers needed theoretical foundations and the ability to use modern EDA tools, Sodini asserted, "it's the hands-on time that makes engineers reasonable people."

Entrepreneur Nicky Lu, chief executive officer of ETRON Technology (Hsinchu, Taiwan), seconded the latter message. "In my companies, we have to train newly hired graduates a lot to make them useful. We get digital designers who can't calculate the drive necessary on a net from the fan-out without going back to the simulator. We get RF designers who can't do layout, and RF designers who learned 200-MHz techniques in CMOS when we are working at 2 GHz in GaAs. And we find that new grads have an increasingly narrow skill base."

Lu struck a resonant chord in the panel with another observation: "We see graduates who are too much dependent on simulation instead of thinking things through."

Hugo De Man of Katholieke Universiteit, Leuven, Belgium, agreed with Lu. "We have kids coming into college who are already using PCs. The first thing we need to do is take the computer away from them, and make them solve problems with a pencil, paper and their heads."

Additional support was readily available from National Semiconductor analog guru, industry curmudgeon and renowned SPICE critic Bob Pease. "Have these students learned to design something without SPICE? Do they understand its limitations before they start using it?" Pease raised the question of whether, when an engineer gets virtually all his training on simulation models and almost none on the bench, we are training a creative engineer or merely a software operator. "My favorite programming language is a soldering iron," Pease said.

Lu and De Man agreed on another point as well. "I see new hires come in, and they just sit down with their computer and never talk to other engineers," Lu said. "We need to teach them to work in teams, not in isolation. In industry, you can't get anything done by yourself."

De Man took a different cut at the same issue. "We have to make students from the beginning work in teams, not off by themselves in a corner. But the problem doesn't begin with the students. It begins with the professor, who is locked into a system where if he can make something run three percent better than the paper at last year's ISSCC, he gets cited. Three percent better than last year isn't a system. We have to change the system that discourages teams, and get both the professors and the students to work in teams.