Quote:
Originally Posted by grant
I think the hardware is probably designed well enough. It's more likely that the drivers are bad. That's why it's important to release detailed specs and good docs so that drivers can be written for _any_ system or rewritten if the manufacturer's drivers suck.
|
The problem with so much hardware design (and I worked on enough for the government to appreciate why) is that the drivers are part of the hardware design. Unfortunately too many hardware designers don't understand software and too many software writers don't have a clue about what is really happening in hardware.
I was often requested to use both software and hardware debugging tools in conjunction - be it to align disk drives in the old days or solve a bus problem in the later years. I would end up writing my own diagnostics because management had never bothered to coordinate hardware and software people in developing those necessary tools (that cost money). In many cases, communication was so poor between the hardware designer and his software assistant (who was writing the drivers) that some hardware and software departments all but would not talk to one another. Another tactic was no documentation for the hardware / software interface. Quality was not as important as cost controls and schedules.
Therefore hardware caused strange problems where it does not affect the manufacturer's bottom line - with you.
USB was a unique and rather unusually reliable creation. To address this hardware/driver problem, the USB consortium would sponsor 'plug-fests'. Then when the product did not work with so many other USB products, the boss probably could attack both his hardware and software people to either force a team or create open warfare.
From the user's perspective (if he does more than just shotgun or swap components), the hardware (chips and software/firmware) incompatibility was often reason for strange failures.
Today, bad hardware causes Windows systems to not recover from standby or from hibernate. Incompatiblity causing clone computer board manufacturers to have long lines testing their boards to video cards from virtually every video card manufacturer. And still I have seen at least one motherboard that was not compatible with a large number of AGP video cards - literally dropping pixels when run in 4X mode. Again, defective hardware. A new version of the same board appeared but months later meaning the manufacturer knew of the intermittent and was designing a replacement while still selling that marginal and earlier version. Again - hardware failure. You don't matter when it is not a brand name product.
|