美国工程师点评苹果天线门的责任

Apple iPhone 4: Can't engineers just say no?

by Karen Field

Apple's recent troubles raise questions about who exactly is empowered to stop-ship when a product doesn't conform to design requirements

Howard Fidel, an analog design engineer at Schick Technologies, a manufacturer of digital dental radiography equipment, knows only too well how difficult antenna design can be--his current project involves a Wi-Fi dental device with a diabolical number of use scenarios. The product line Fidel supports has a five- to ten-year development cycle with approximately 1,000 units shipping out per month. He says he can only imagine the pressure placed on engineers in the consumer electronics industry and Apple in particular.

"Given the rapid product life cycles and Apple's secrecy, engineers probably just weren't able to do adequate beta testing. They probably gave it to a few engineers to test and then just blew a million units out of there in a month," he says.

Other engineers seem to be in agreement: In a recent flash survey by EE Times asking engineers what they think is going on with the iPhone 4, the most popular theory proposed by 37% of the 220 respondents: "That's what happens when engineers are pressured to get a product out the door.

"Only the people who were on the project and in the communication loop at the time really know what happened," says Henry Martinez, senior vp for a software development company. Speaking on his new iPhone 4—he's had minimal problems with dropped calls –he wondered whether the data was sufficient to be considered a risk, and whether engineers at Apple were empowered to pull the emergency stop switch, like workers on an automotive assembly line.

He notes that when engineers do raise red flags, it's critical to sort out whether those concerns are simply the boy crying wolf or the true canary in the coal mine.

Antenna designer Fidel says that he doubts that since Apple engineers knew about the problem they would have let the product ship without management signing off on it. A widely-circulated article reported that a senior engineer at Apple raised concerns about the antenna's design.

Though Fidel notes that he has not delivered a design that he felt was deficient in the last 20 years, it did happen once early on in his career. "When I was first out of college and working for a startup company, the president had me ship product that we knew did not work properly," he says. "It was a strategic decision based on the lesser of two evils: being late or being wrong. Since it was a low-volume product we were able to easily fix everything later. He adds that he believes that it's every engineer's responsibility to document deficiencies, and then it's up to management to decide whether or not a delay is acceptable.

Martinez speculates that sometimes engineers may be reluctant to voice their concerns because of misguided job-security fears. "They worry that blowing the whistle might get them in trouble, when actually not doing anything creates more problems, because any defects will be discovered in the field, instead of in testing," he says. "Which everybody knows is much more expensive in the long run."

Alan Wu, a hardware validation engineer at AMD, says he was surprised that a defect so seemingly obvious wasn't caught before Apple shipped the product. "I don't know about the culture at Apple, but something like this should have triggered a stop-ship, it shouldn't have been let go to production. Product development is always a balance among cost/performance/risk. Apple clearly made a risk call and it backfired on them. There are probably a few people there right now saying, ‘I told you so.'"

In general, Wu says that engineers should report every concern so that if anything happens, at the very least they can say they reported the issue and were told to ship the product anyway. "By doing so, the responsibility is passed on to management and/or more senior engineers to make that final decision," he says.

Not only that: As Fidel points out, "Engineers tend to want to perfect everything, and some projects would never get finished if management didn't step in.