Reading uneducated users opinions about signal meters ("phone A is better because it has one more bar than phone B") gives me that same sad feeling about humanity that talking about politics with "low information voters" does. It makes me wonder how stupid I am in areas that I don't understand very well (say, chemistry or medicine).
I've had the fairly unusual experience of getting to sit in meetings about controlling the display of signal bars. It's interesting because you're doing a bunch of fairly complex math on real phenomena (signal level, noise floor, corrupt symbols or low level frames, and so on) to turn it into one of six messages to a human being, without any words. If you deflate the value (too few "bars") you get dissatisfied customers -- they think that your service sucks because they've never seen more than three bars, even on a sunny day. On the other hand, if you inflate the number, you end up with a lot of support calls -- people that have five bars and bad voice quality or slow performance.
Basically, I gathered that the messages that the bars convey to the user are:
0) The device can't see the tower or would be unable to place a call/data connection.
1-2) The device is on the network and believes it can perform but the user should expect degraded performance or connection failures.
3) The connection should be OK and performance should be up to advertised specification (full speed, no drops), but you aren't getting the best signal possible, and should pay attention as it could drop from here.
4-5) Your signal is very healthy and you shouldn't expect anything better than this. It most certainly won't be a problem for you.
I'm kind of surprised, given the marketing push for "bars" in the cell industry that we don't have 7+ bar signal displays these days -- it would be a pretty safe way to give more bars without actually compromising the underlying message. I was pleasantly surprised when I saw an iPod with a battery meter that was calculated to the pixel (a smooth bar, not three simulated LCD segments) -- it removes the ability to complain about being down "a bar" after listening to M songs when someone used to get N songs -- completely unscientific but the basis for tons of bad press about battery life. I'm surprised no one has tried this on mobile signal (that I know of).
Wow... I would so much rather just have (perhaps a subset of) the actual data always on my display. I don't have an iPhone, but there's some room on my little screen that could show -90 when I'm in a cave and, like, -9 most of the time (or whatever it may be).
Seriously, people, are we supposed to be so dumb that we can't do that kind of judgment call on our own? I thought that that kind of thing was what the human mind was supposed to be good at and that computers are supposed to be bad at, yet here we are having a phone do it for us?
A very readable article about what goes into the bars on your cell phone. I liked it especially because it doesn't jump on the "OMG ATT/VERIZON/WHOMEVER IS THE DEVIL" bandwagon that seems so popular on the internet.
the map is not the territory. anytime you have people associate a positive with some arbitrary measure, the people controlling that measure have an incentive to inflate it. this makes it appear that they are improving things while doing nothing.
see: the economy, public schooling, housing, etc.
I've had the fairly unusual experience of getting to sit in meetings about controlling the display of signal bars. It's interesting because you're doing a bunch of fairly complex math on real phenomena (signal level, noise floor, corrupt symbols or low level frames, and so on) to turn it into one of six messages to a human being, without any words. If you deflate the value (too few "bars") you get dissatisfied customers -- they think that your service sucks because they've never seen more than three bars, even on a sunny day. On the other hand, if you inflate the number, you end up with a lot of support calls -- people that have five bars and bad voice quality or slow performance.
Basically, I gathered that the messages that the bars convey to the user are:
0) The device can't see the tower or would be unable to place a call/data connection.
1-2) The device is on the network and believes it can perform but the user should expect degraded performance or connection failures.
3) The connection should be OK and performance should be up to advertised specification (full speed, no drops), but you aren't getting the best signal possible, and should pay attention as it could drop from here.
4-5) Your signal is very healthy and you shouldn't expect anything better than this. It most certainly won't be a problem for you.
I'm kind of surprised, given the marketing push for "bars" in the cell industry that we don't have 7+ bar signal displays these days -- it would be a pretty safe way to give more bars without actually compromising the underlying message. I was pleasantly surprised when I saw an iPod with a battery meter that was calculated to the pixel (a smooth bar, not three simulated LCD segments) -- it removes the ability to complain about being down "a bar" after listening to M songs when someone used to get N songs -- completely unscientific but the basis for tons of bad press about battery life. I'm surprised no one has tried this on mobile signal (that I know of).