links for 2010-09-14

  • "Let's take out the second pilot. Let the bloody computer fly it," O'Leary told Bloomberg Businessweek, adding that a flight attendant on each flight could be trained to help land a plane if something goes wrong.
    "If the pilot has an emergency, he rings the bell, he calls her in," O'Leary told the magazine. "She could take over." <– Um, what happens if the pilot doesn't get to ring the bell? Boy, I'm glad this guy runs an airline…
  • "Snapshots are not backups. As the snapshot file is only a change log of the original virtual disk, do not rely upon it as a direct backup process. The virtual machine is running on the most current snapshot, not the original vmdk disk files." I like this document, mainly because it's concise and sums up what I've been doing for a long time. I think I'll make it doctrine where I work…
  • Ha, 19/20 — I screwed the "Windows software running on Linux" up, didn't catch what they were asking.

3 thoughts on “links for 2010-09-14”

  1. I think you lack understanding of just how good autopilots are these days. Most of the time the pilots act as supervisors for the autopilot and not much else, even during landing, which has long been considered the most difficult part of piloting an aircraft.

    Pilots are quickly becoming redundant and it really won’t be long before computers take over the job entirely. That’s one of the reasons I didn’t bother becoming a commercial pilot when I was 17 – I could already read the writing on the wall.

  2. That’s for sure, I had no idea autopilots were capable of that. Just looked it up on Wikipedia, a little scary, in my opinion. I know there’s human error, but I don’t trust programmers, either.

  3. Most of the time, the pilot is following a checklist to ensure that they’re doing things right and nothing gets forgotten.

    You and I like to refer to checklists as “shell scripts”. 🙂

    Pilots rely on the same instruments that the onboard computer uses – in fact, the onboard computer in the vast majority of commercial planes is what is *giving* the pilot the readings on those instruments in the first place. Should the computer or the measurements from the instruments (to which the computer is connected) be wrong for any reason at all, the pilot is no better off than the computer. A pilot can’t possibly guess the airspeed of his aircraft by looking out the window should the pitot tube ice over or get a rock stuck in it.

    As for “What happens if something unexpected happens, like that emergency landing in the Hudson river?” Well to be honest, a computer probably would have handled it even better. Again, emergency, unpowered landings have checklists too. In this particular case the pilot did not actually have time to complete the checklist due to the low altitude at the time of the incident, and missed the part that would have kept the aircraft afloat much longer than it was. That aside, the computer can figure out that the nearest airport is *not* within reach much faster than any human could, by virtue of computing its distance compared to the aircraft’s rate of descent. The decision for a water landing and the proximity of the river is likewise an easy one, and is probably even on the pilot’s checklist.

Comments are closed.