Was FORTRAN buggy 4338

On Wed, 20 Sep 06 12:29:17 GMT

True - and even then you won't catch them all. Think of the one that only happens after 40 years use at double the design usage and with poor maintenance. What you can do is first cover every failure you can anticipate *before* you build any hardware and add to that any unanticipated failure you ever see with real hardware and supplement that with any failure modes that those real failures cause you to imagine.

You don't - and it doesn't. What it does need to do is appear to the software as though it were the real thing and that includes presenting the same garbage under failure conditions. Noting that verifying that a real aircraft produces the same garbage as the emulator under the failure condition that a wing has just sheared off along a particular line while doing mach 0.8 at 35000 feet is a trifle expensive and difficult to achieve. Having the real aircraft available does not help here.

I would be surprised if it were not the case that the software for a new aircraft is usually available and has been tested in emulation and some level of hardware emulation long before any aircraft are built.

Part of this is because with modern hardware the emulator always comes first and the hardware design is sanity checked in simulation before it is ever made as hardware. If the resultant hardware does not match the simulation then the reason for the mismatch should be tracked down and repaired.

-- The computer obeys and wins. A better way to focus the sun You lose and Bill collects. licences available see