alarming aliteration

steadily spreading stultifying pseudo-statistical servitude. ;)

Cheers

Phil Hobbs

-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics

hobbs at electrooptical dot net

Speedy slam, sir.

Six-sigma, by itself, seems to be a good idea. But like a lot of other things, once it falls into the hands of petty bureaucrats, it can get pretty grim.

Or perhaps at least continuous process improvement.

Savoring suspension of sensible system successes.

(Trust me, I know about this stuff.)

Stunts so psychologically stupefying, several sixers slept silently

ChesterW

We have shipped over 2200 boxes to Corp X in the last three years, with one field failure, both center pins broken off a pair of male BNC connectors.

They gave us an F on our quality audit, and a B- for engineering, for an overall D- score. We don't have enough 5whys or 6 sigmas or 5WH2's or 8Ds or Pareto Charts or responsibility matrices. We don't have anyone in charge of labor relations or anyone responsible for sustainability and conflict minerals.

Meanwhile, they have 11 real quality problems that nobody is doing anything about.

The problem with the continual improvement idea is that eventually you get down to the shot-noise dominated regime--your underlying failure rate may have gone from 0.5 per year to 0.2 per year, but how are you going to measure that on an annual basis? And of course in that unlucky year, if the Quality Commissars have their way, you're going to get fired.

Cheers

Phil Hobbs

I've always found it weird how many people have their eye on something othe r than the ball, and pursue it regardless of what actually matters. And if you dare put what actually matters above their petty concerns, you get that F. Its one of the things I don't like about large corporations, they tend to fall into that game.

The good news is that over time they either get much more efficient, or win d up out of business - normally the latter. Its happened to so many you'd t hink more would learn the lesson.

NT

If you replaced ISO-9000 with Six-Sigma, I'd probably be right with you but Six-Sigma is an inherently dumb program.

Suppose my board has a 7805 regulator. We can measure its output!

What we'd need to do is measure the 5 volt rail on every unit and do Statistical Process Control. If we see a trend, we need to hold meetings.

(The example about filling cereal boxes is especially absurd. Cams and pulleys!)

You suffering son of a Siberian sheep shipper!