The Magical Number Seven
The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information
An interesting reference to Human Interface Design. Did I mention I love this shit? I remember explaining this to someone who thought I was a nut job driver. My arguement was that although we have 5 senses, we still can have several processes going on in our heads at any one moment, and the number of processes wasn't limited to by the number of senses one has. My examples of inputs/monitoring processes while driving were Rear Car Count, Forward Traffic Pattern, Steering wheel vibration, accelerator pedal vibration, seat vibration, tire traction sound pitch, and it was which mode of driving you were in that would dictate which processes to follow.
Did I mention I hate being called a crazy driver?
posted by Chief Jimbo at 4:54 PM
3 Comments:
Whoever told you the number of things you can process is limited to the number of senses is retarded.
Although, usually, with a thing as complex as driving, you aren't simultaneous processing all variables. You're only conscious of a handful of stimuli at any one time, and you're constantly shifting your attention to accomodate everything. You aren't aware of all things at the same time; your attention is devoted to each item for however minute a moment, and each little bit of info you pick up you're applying to your concept of the situation as a whole.
In any case, the article was interesting, but I didn't think they addressed the effect of variation between individual items in an absolute judgment. For instance, pitches (or colors or tastes) are easier to distinguish when there's a greater relative difference between each individual item in the selection. Like, picking between black and white is easier than picking between gray and ever-so-slightly-more-gray.
That, and I though it was kooky that they expressed the capacities as "bits," but then used decimal precision (eg, 2.3 bits). Bits are on or off, stupids.
7/12/04 4:04 PM
I think the author really screwed the pooch when he decided to use the term "bit" although when first introduced, it made enough sense. (Us mechanical types think of bits like "drill bits" "ball end mill bits") I guess he stays pretty consistant in the actual meaning of it, but it added a level of confusion.
The bit was described as a yes/no condition. "Is this man above 6' or not?" I think my confusion comes because he tries too hard to BE binary. Upon first read, I was like "Oh, so they get to identify them with roman numerals and then have a sheet in front of them that has a check list for each tone they hear. For each tone they have to put down "yes" or "no" for _every_ roman numeral on their check sheet. 'This is not IV, not V, it is II'" If this were the case, 2.5 bits would mean they would only be able to discern 2.5 different tones, which was shown not to be the case. The case was that 2.5 bits the "tone channel" for humans was 2.5 bits, allowing for 2^2.5 (~6) pitches to be discerned from. So essentially, any given tone in this experiment could be represented as a binary number but since the exponent was not a round number, it wouldn't be portrayed in the usual 8_4_2_1 place holder fashion (0101 = 5). So in this case, it would be a 2^2 binary representation (4_2_1) with a range of 000 to 110 with each binary number inbetween representing what the subject deemed a roman numeral. When he goes into bits and chunks, it makes a lot more sense.
In terms of variation, I think this did a good job of determining the upper normal limit of absolute judgement in certain cases (tones, colors, etc) and explained dimensions pretty well. In terms of black and white vs gray and grayer, they are all on the same one dimensional scale. A good test to see the channel bandwidth in this case would be 5 equally seperated black/white/gray mixtures. I think what you might be looking for is channel resolution: how well can someone make an absolute judgement when the choices are so close together.
I guess I wasn't stating that a good driver takes in ALL of these inputs and processes them ALL at the same time. It was more of a situational process scheme. There are of course core monitoring processes (such as relative velocity and immediate threat prioritization) but the situation dictates which inputs to monitor. My theory is that you can monitor different processes, but the amount processes being focused on, and the focus on which process, would dictate how much accuracy you would get from the monitoring processes you aren't intently focused on. That's why you can still talk on your cell phone while maintaining some level of driving ability but in doing so you have a higher chance of rear-ending a 93 red honda prelude.
9/12/04 5:54 PM
I had to re-read your response like, six times. You're too smart for your own good.
And you're right, it appears I was talking about channel resolution, if that's your description of it (not familiar with the term). Basically whether or not people'd make more mistakes if the items they were attempting to distinguish between were relativley more similar.
And now, for your enjoyment, the hand trumpet.
Pffft Pfft Pfffft Pffft
9/12/04 6:22 PM
Post a Comment
<< Home