Sunday, April 13, 2008

What if all signaling was like your dog?

Our dog has a series of signal methods. If outside:
  1. Stare at door
  2. Sigh
  3. Scratch lightly
  4. Scratch firmly
  5. Bark
These are pretty easy to interpret, as she pretty much always wants in. Occasionally its a notice "there's a possum" but we don't care, so that's fine, and we want her in so the net effect is the same. But indoor signaling, or rather "proximity to humans" signaling is different. The methods, again in increasing order, are:
  1. Stand nearby
  2. Walk away (leading you somewhere, but we never notice)
  3. Stare longingly
  4. Stare pathetically
  5. Rest head on you
  6. Bring you something
  7. Paw you (or scratch at things of interest)
  8. Jump up on you
  9. Leap around
  10. Yipping or wooing
  11. Talk
  12. Bark
But in this case, it could mean /anything/. Food, water, walk, play. Anything. And, therefore, its pretty easy to misinterpret. We routinely have to disregard the messaging as we know she's got food and water, and just came in, so doesn't need anything. Sit on my lap and shush. Exceptions are even worse. One night she just bugged the hell out of us for a while at 4 am. It only went to stage 4 or 5, so we disregarded (not critical enough, if its like "I must go outside") to get up. She can hold it. Turns out the pickup truck was being stolen. Well how the hell was I supposed to get "the truck is being stolen" from that? Its fun to imagine atrocious computer systems that have only one signal, for notifications, errors, everything, in much the same way.


iamyou said...

They have them... they are called post codes.

Steven Hoober said...

When I last had a computer that beeped at me on startup, most of the error conditions that were signalled were unique. Requires interpretation, but unique so decipherable. Like, no video (and not set to what was then referred to as headless operation) gives 3 beeps.

A related topic is technically decipherable, but functionally useless signalling. Like many alarm codes (nuke plants, some aircraft) where they just keep stacking alarms, so that none but the most common are memorized, and it's just a wall of sound.

Or, conditional coding. On aside from hypoxia effects, it is likely that the warning horn for loss of cabin pressure was confused with the more commonly encountered identical signal for improper takeoff configuration. This is especially egregious as one is procedural, which can lead to danger, while the other is an actual immediate warning.