It's fitting that the longest, most drawn-out Semicon West ever took place during what has become the longest, most grueling industry downturn ever. I haven't spoken with anyone who wants to endure another Semicon with the bass-ackwards scheduling of this year's event, where the San Jose back-end portion preceded the San Fran front-end days with a weekend in between. Likewise, I don't know anyone who wishes the downturn would linger longer.

The overriding theme at West '02 was the economy (duh) and as far as I can tell, that's still the foremost concern on chip people's minds. The show may be fading into memory, but my retort to folks asking me when the upturn might happen is the same now as it was in mid-July: My visibility extends to the end of the day.

The wounds are still healing from the beating that publicly traded SEMI members and their customers took in the stock market during the show weeks. A check of the indices for mid-late July reveals that the bear took its biggest bite out of market caps as attendees window-shopped for tools or materials and exhibitors put on their game faces.

The Dow Jones Industrial Average bottomed on July 23 near 7600, down a mere 1800 points or 19% from earlier in the month. The Nasdaq average plummeted from just over 1397 on the first day of Semicon (July 17) to a close of 1229 and change on July 23 (a drop of 12% in less than a week). In fact, the tech-heavy exchange went as low as 1192 that day—its 52-week nadir—before rebounding at the closing bell. The following day, it started north again, gaining about 51 points, but bottomed out almost two weeks later, closing at 1206 on August 5. Since then, stock prices have fared a bit better, so at least the most serious pounding has abated.

It was the only Semicon West I can recall (and I've now attended 15) where there were no fresh outbreaks of allegedly paradigm-shifting technologies. Sure, copper and low-k dielectrics, SOI, 300-mm fabs, and advanced process and equipment control are still moving forward into their second-, third-, and umpteenth-generation versions. Atomic layer deposition and high-k dielectrics are edging away from the exotic realm and toward the mainstream. The metrology equipment crowd is getting closer to the holy grail of achieving high throughputs and extraordinary sensitivity at the same time on the same tool. The greening of cleaning technologies and chemistries also has made significant progress, where processes use less water and fewer toxics and the wafers are cleaner than ever.

But a definite shift in the marketing language played out at this year's show. For the marketeers, the use of the "nano" suffix was rampant, with a losing battle being fought over where "micro" ends and "nano" begins. (For the record, Intel's scientists define nanotechnology as devices with minimum feature sizes of less than or equal to 30 nm, so why not give up and let them have their way—you know they will get it eventually.) Most companies are making the transition from microns to nanometers, with geometries or feature sizes of 0.13 or 0.10 micron or less converted to 130 or 100 nanometers or less. As the "nm" abbreviation replaces the "µm," one less Greek letter to edit sounds fine to me.

The rush to nano-ize the jargon reminds me of when the contamination control marketing folks got their mitts on the suffixes and modifiers preceding "pure" and "clean" in the late '80s and early '90s. Clean begat "very clean," which begat "ultraclean" and its offshoot, "hyperclean." Pure sired "high purity," which inevitably fathered "ultra high purity." The exact level of cleanliness or purity denoted by these terms was rarely precise, so one person might define "ultra high purity" at 500 ppb while another said it was 100 ppb. Problems arose when ostensibly technically oriented people started throwing the terms around without really explaining what they meant.

Recent news stories about major 130-nm process problems, with vias failing between copper interconnects, suggest there were more than economic woes being discussed at this year's show. But then failure modes aren't exactly what an industry on its heels wants to parade as the next big thing.