On Mon, 19 Dec 2005 10:05:59 GMT, Roedy Green <[EMAIL PROTECTED]> wrote:
>On Sun, 18 Dec 2005 18:42:52 -0800, robic0 wrote, quoted or indirectly >quoted someone who said : > >>If the software opens a file and is in the middle of writing to it, >>then the user dumps the power to the machine and ends up having to >>reformat, thereby losing all his data, at what point does the >>liability stop? And how is fault proven or dished out? Does the >>law specifically state "repeatability" in its language? > >It would expect it to work much the way a car works. If you have an >accident, that is your fault. If the fuel pump is badly designed so it >catches fire, that in the manufacturers fault. You'ld have to prove the fuel pump caused your accident wouldn't you? I'm reversed when it comes to engineering. I always assume defects when buss loads of people are killed. If software ever guards lives that isin't certified then its a manufacturing defect. That is imbedded software though. Not the for public consumption. I know that fly-by-wire military software has 100 levels of precaution. Hey but its a 7 million dollar plane and a 700 billion dollar budget. The written requirements for a single design is a book 5 inches thick. Ever see that for Joe bullshit software designer? -- http://mail.python.org/mailman/listinfo/python-list