Did anyone check the iPhone bar changes back when Apple increased them the first time?
Apple's new bars are still more optimistic than some phones. For instance, they show 2 bars where a Blackberry would show 1. And they show 3 bars where a Blackberry would show 2. They're more in sync at 4 and 5 bars.
Signal bars are mostly meaningless for WCDMA-3G anyway. The overall noise floor is much more important than the pilot channel signal strength.
So why use bars at all? Some have suggested Yes-No flags. I think the reason is that it provides at least some feedback on signal-vs-location. For good example, if you move around a little and make a call from a much higher signal reception area, it's probably a good thing for you and other users of that cell, since your phone might not have to artifically raise the noise level for no good reason.
+1 - hallelujah! Somebody gets it.
What we need is an indicator of performance from the end-user's perspective. The rate at which data is being successfully communicated would be a good possibility (I note that for MacOS X one can see the rate at which data flows to and from a network). Another possibility would be the ratio of the signal to the noise (S/N ratio, SNR). Or better yet, a reading that indicates the empirically derived probability (from real user data from volunteers) of a dropped call.
The only way Apple can weasel its way out of this mess is pure, unadulterated honesty and transparency, and the fix for the display of bars smacks of nonsense (if you believe it, perhaps you'd be interested in Spinal Tap's '#11' guitar). I hope Apple will say, 'Yes, we fixed this minor bars issue, which we explicitly recognize does not change the quality of reception, but we are working to optimize the software and future hardware designs to maximize reception'.