On Sat, Apr 18, 2009 at 10:19:12AM +0100, mike wilson scripsit: [snip] > Bill and I aren't really having intermittent problems. My own feeling > (and it's nothing more than that) is that it's a luck-of-the-draw thing.
Some, yes, but that's typically got a lot to do with what happened to the part after it left the factory. > The motherboard is a complex concatenation of even more complex > subassemblies. We all know the possibly apochryphal story that all > processors come off the same assembly line and that it's only testing > that defines which has the fewest faults and is therefore the best. That's effectively true; testing determines clock speed ratings. > Put two of those hyper-complex units together and you are bound to > have a huge range of interactive results (faults), many of which will > have no visible effect. Well, no; this is one of the good things about digital, you can define the interfaces relatively independently of the processing, and you can treat anything bigger than value <foo> as a one. It makes a modular, black-box subassembly approach to complexity management much more workable than it is with mechanical parts. > Put a number of them together in their own complex environment and, > personally, I think you are exceptionally lucky if there are > absolutely no visible problems. I would happily bet that, if you > examined it closely enough, anyone's computer would not match exactly > the manufacturer's specification. The specification always includes ranges. :) It's quite possible to get something that matches. The problem is more usually differential failure under load (it was fine until you used it hard for awhile), and specing components to be more reliable in that respect does cost more. (And the big box guys generally don't do it.) -- Graydon -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net to UNSUBSCRIBE from the PDML, please visit the link directly above and follow the directions.
