7/11/03 7:59:14 PM, Alex Kilpatrick wrote: >> We don't consider a design as working until 100 units pass >> the production line and into the field without hands being held. >> > >And that is perfectly sensible for a production product. At the risk of >being redundant, this is a *research proof of concept*. If it works for >the duration of a demo, I'm ecstatic. If the funding agency is >interested, then they will take the device and send it off to someone >else for real "productization," which could conceivably (probably >likely) involve a complete redesign. I'm a researcher, and don't get >involved in production issues. I agree they are very important, just >not to what I am trying to do. And personally, I enjoy making the first >few; after that I am not interested. The trouble here is that the sort of design sloppiness that results in a prototype working on a test bench while the production units fail in the field *also* can result in a prototype working on your test bench but *not* in the hands of the client or the funding agency. The problem isn't just excessive variation from unit to unit (which is more a problem of failing to design for manufacturability), it's individual units' lack of robustness to variations in their environment. Whenever you design something on the assumption that Mr. Murphy won't pay a visit, you're actually buying him a first-class airline ticket and a room at the nearest 5-star hotel. -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics