On Thu, Mar 02, 2006 at 08:26:10AM +1000, Rob Studdert wrote: > On 1 Mar 2006 at 10:24, K.Takeshita wrote: > > > Adding to this, Kyocra actually approached Pentax for help when they > > (Kyocera) had been plagued by the noise problem (even at ISO100), as Pentax > > had solved it at the time. But Pentax respectfully declined. So, Kyocera > > was on their own and got back to the drawing board. > > Pentax MZ-D was truly ready for the production (some production models are > > actually being used within Pentax) but as Bob said, the price killed it. > > Thanks for the additional info Ken. So in essence where Pentax screwed up was > in securing a solid pricing and procurement deal? Having being in design I > find > it hard to understand why anyone would sink that much R&D into a product when > the pricing of one of the most fundamental of the components wasn't pre- > negotiated. Surely the CCD pricing would have been a factor for the project > being given the nod in the first place, surly it would have had to have > theoretically been proven a viable product before any prototypes were > produced?
It was a viable product, *IF* the sensor had been made available, on the projected delivery date, at the estimated price. If you think a company the size of Philips is going to give guaranteed pricing and delivery dates for a product that isn't yet in production you're not living in the same world as I am - you can't even get those kind of deals when the relative sizes of the two companies is reversed when it comes to the cutting edge of design. I've seen that (from the other side) when I was working at SGI. Sure, we could get pre-agreed prices from the chip fabricators for putting the design through the foundry, and we could even get some pretty good estimates on yield problems caused by silicon impurities. But as far as other defect estimates went? Forget it. And, in my experience, any cutting edge foundry will come online late, and run into unanticipated problems (such as trying to solder connections closer than ever before). And the best design simulation and verification tools don't help you when the problem turns out to be electrical and/or thermal noise on the chip; it's not a design issue, it's an implementation problem. You think Pentax "screwed up". I think they took a risk, gambled on Philips being able to deliver the chip, and lost. In hindsight it looks like a bad decision, but presumably somebody thought the potential gain was worth taking the risk - after all, if Pentax had got that camera to market when they hoped what would their market share be today? Nobody knows, but I'd bet it would be much larger. As for gambling on future chip availability - how many people have bet their company on Intel, only to see that *they* can't deliver chips on their projected timeline, either.
