As design and development engineers and managers, you've changed, too. You're more business-oriented, market-focused and team-oriented than in 1992, though engineers had already started changing in the '80s as corporations expected more than technology from them. Fundamentally, you remain a conservative integrator of new technologies. You use what works, what's proven, what's affordable and what can be tested.

We asked you what was working out there in terms of new technologies.

• Linux-35 percent of you say it "works well." This open, Unix-derived language has been picking up steam since its first official version debuted in 1991.
• System-on-chip-31 percent express satisfaction with SoC technology, though 7 percent have found it unsatisfactory in solving design problems.
• Formal verification-A quarter of you think it works well, 11 percent don't like it and 63 percent haven't tried it.
• Network processors-21 percent give them a thumbs-up, with only 4 percent sour on these dedicated processors. Three-quarters have never tried them, and if you're not working on a communications networking system, chances are you never will be exposed directly to them.
• Streaming media-While 20 percent of you approve of this technology, 16 percent do not. That's our biggest negative rating.
• Web-based design tools-19 percent like it, 14 percent do not.
• Silicon IP-17 percent say IP cores work well. See more detailed information later in this chapter.
• XML and open scripting languages-17 percent approve, 4 percent do not.
• Reconfigurable processors-16 percent say they work well, 5 percent call them unsatisfactory.
• Reconfigurable communications switches-12 percent like them, 2 percent do not.
• Embedded Java-11 percent approve, 8 percent don't.
• Bluetooth-Only 12 percent of our sample is familiar with this communications protocol; 8 percent who have used it say it works well.
• Direct Rambus memory-A split vote here; 8 percent approve, 8 percent don't.

You can do the math.

Anywhere from 60 percent to 85 percent of those who responded to our survey replied "haven't tried it" to questions about system-on-chip technology, Linux, network processors and silicon intellectual property. In light of these results, you might well ask, "what are these guys doing out there?"

The answer lies in the fact that EEs use only a fraction of the technology available to them at any one time. A system-on-chip designer doesn't necessarily need XML and open scripting languages. A user of embedded Java may not have tried streaming media or direct Rambus memory. So it isn't the same 60 percent of you who's saying you haven't deployed it.

We also have to consider that 14 percent of our sample are managers-who may or may not actually get their hands dirty in the technology mosh pit-and 1.5 percent are corporate managers who most certainly do not. Close to 9 percent of you are software designers, while 4 percent are component and chip designers, who are exposed to a narrower spectrum of new technologies than, say, systems designers. A state-of-the-art processor designer may well check off "haven't tried" 10 of the 13 new technologies posted here.

So what was "hot" at the beginning of the Clinton years? And have these technologies panned out? In the case of high-definition television (HDTV), it hasn't-yet. Back in 1992, some 94 percent of the respondents said HDTV was "on the rise," the highest-ranked hot technology. Needless to say, it's been a very slow, torturous rise. Only in the last year have the first HDTV sets appeared in retail stores, carrying inflated price tags of $4,000 and higher. HDTV ran into a series of protocol, governmental and technical set-backs, but it finally does seem poised to make inroads into the television industry. The next "hottest" technology of 1992 was 3.3-V technology. Been there, done that. We got there, and kept going down to 1.8 V and lower. The picture is more mixed for neural networks and fuzzy logic, selected as on the rise by 90 percent of our 1992 engineers. You can find both technologies embedded in systems today, especially the so-called "smart" technologies. Whether you could term their acceptance as "hot" is a matter of opinion.

However, there's no doubt about the next technology, tapped as promising by 88 percent of our 1992 sample. Palmtop computing is here, and it's big, albeit perhaps a bit later on the market than our 1992 engineers might have thought would be the case.

The next two made it, too. Multimedia, selected by 84 percent in 1992, is an accepted part of the technology scene today, as is parallel processing, picked by 80 percent. Interestingly, optoelectronics is just now putting down some tread as networks increasingly cross over to fiber, yet one-third of the 1992 respondents thought of it as "mature."

Intellectual property cores weren't around in 1992, when President Clinton took office, but they've become a prominent option for system designers.

"For industry-standard IP such as PCI and VGA, our engineers cannot add value-it is simply a check box," a manager writes. "Use of IP cores allows us to essentially buy compliance and allows the engineers to use their creativity and expertise on features that differentiate our product."

Since 1999, the percentage of respondents who are involved in incorporating outside intellectual property into a design has risen from 20 percent to 30 percent, a significant increase.

What skills do you tote in your mental toolbox? About half of you remain firmly in the hardware camp. Another third of you say your work requires both hardware and software skills, while 17 percent work primarily in software.

But what really stumps our readers are people and business skills:

Here are some of the biggest hurdles cited by our readers:

• Product requirement stability, especially during high turnover in product management and marketing areas. "Engineering frequently is chasing a moving target and is blamed for nondelivery and late delivery."-senior engineer
• Project management. "Like most things, takes experience to get a feel for it."-senior engineer
• Scheduling. "A million-gate ASIC took nine months to convert from one vendor to another instead of the three months scheduled."-senior engineer
  

See related chart