Nobody expected Symbian to survive much longer anymore, but going Microsoft was the dumbest option.

The largest hardware market share with the worst possible OS. Could only come from an American mindset CEO: "Hey, look! An ultra fast way to make the Windows Mobile market share go up and get fat bonuses quick! Let's show these fancy Europeans how a real OS should be!"

Why didn't they go Android then?

Nokia fanboys (like myself) are there for the intrinsic cool factor of a Nokia phone (no matter if cold numbers point to better hardware or more modern OSs out there). They just needed to keep the flame alive sprinkling some edgy tech here and there, no need to sell their souls to the devil. Apple lives up on coolness alone.

What a way to completely "finnish" up with the coolness...

Joao S Veiga

Please excuse me if this sounds pedantic but it is relevant for the subject.

When we're talking about levels of -51, -91, -113, etc., these are probably GSM signal levels, and the unit should be dBm.

dB alone is just a ratio, which is suitable for indicating for example Ec/Io, which is sort of signal to noise ratio when in WCDMA mode (the iPhone, being an UMTS phone can use GSM or WCDMA). -113, -51 do not look like Ec/Io values (these are good around -10dB).

When in WCDMA mode, Ec/Io is much more relevant than signal level and I believe phones show bars proportional to that. When in GSM mode, signal level is more relevant, and phones show bars proportional to that.

In WCDMA mode, the level is much less important than the phone capacity of extracting meaning out of the noise. You may have a very low signal level, but if the Ec/Io is good, you can place a call. You may have a high signal level, and not be able to place a call if your Ec/Io is poor. This is also true (but not as critical) for GSM mode.

A conclusive test or comparison should not be based only on signal levels, but use a proper GSM/WCDMA phone test equipment, capable of generating different signal levels (GSM) and Ec/Io (WCDMA), then collecting the phone's received Bit Error Rate, or Frame Error Rate. Believing in the level information the phone itself informs and relating that to dropped calls is quite subjective, as it is not a controlled environment, and what the phone informs may be wrong.

In the analog times (AMPS), signal level was all you could measure, and interference just meant the sound would be crappy (and the occasional undesired "group call"). If there was a good signal level, you could place the call (but the quality could be awful, there was no way of knowing).

With digital, bit error rate is what matters, and signal quality is more important than signal level. As long as the received signal level is above the phone's sensitivity (and here is a place where you separate manufacturers who know their RF from the ones who don't), then that's "level enough". It doesn't help to have a lot of signal strength (power) if due to the signal quality (signal/noise) or your hardware/software quality you cannot pull the useful data from it.

In the digital age, it's not the size of your signal which matters, but what you manage to do with it.

Good point, but just to add some more information, here are my two cents:

When an UMTS phone (like the iPhone) is in GSM (and GPRS and EDGE) modes, the signal to noise could be measured by comparing what the receiver gets during its receive timeslot (GSM/GPRS/EDGE use Time Division Multiplex Access - TDMA) against what it gets out of its receiving timeslot. Nevertheless, this is not how the phones measure the signal quality or signal to noise, which as you pointed out is more important.

What is actually considered as a measure of signal quality in GSM is called RxQual (receive quality), which is derived from actual Bit Error Rates (this is part of the GSM/UMTS standard; there is a table relating RxQual to BER in GSM 05.08). The phone reports RxLev (level) and RxQual back to the network for the network to decide about changing the serving cell site (handover). The phone bars however (as far as I know) show something logarithmically related to signal level. This may be not a good indication if you have a LOT of signal, but it's all interference (good RxLev, bad RxQual).

When a UMTS phone is in WCDMA mode, then signal to noise is even more important, but it is measured by the phone by comparing the level it pulls out of the "noise" when using the correct scrambling code against the total received level. Remember that WCDMA uses codes to separate each data "stream", all mixed in the same wideband frequency channel. The more codes in use, the more "noise" you see out of your own code. This measurement is generally called Ec/Io (UMTS standard 3GPP TS 25.133), and is also reported back to the network, with other information, for handover decisions.

I *think* that when in WCDMA mode the phone bars are related to Ec/Io, not actual signal level. If not, they should, as receiving a lot of signal "power" in your antenna when in WCDMA mode does not mean you can get/maintain calls. In GSM mode the "noise" is interference in the GSM frequency, generally coming from (poor) network design allowing channel reuse in cellsites that can "reach" the same spot. In WCDMA, a lot of "noise" comes from normal network usage, so it's a fact of life. Have you heard that CDMA/WCDMA cell coverage "shrinks" with traffic? This is why. You may have good WCDMA/CDMA quality off peak usage hours but "no coverage" at the same spot during peak hours.

All these measurements are ready to use in the phone's UMTS chipset, as they are mandatory by the standards (which by the way are freely downloadable from 3gpp.org), and the standard is strict about how they are measured/calculated - phones have to do it right to pass compliance tests.

Apple (intentionally or not) may have converted them wrong to the bar display (maybe they did not figure out the logarithms right, being newbies in the RF world...), but the measurements could not be wrong, or the phone shouldn't have passed conformity tests.

and UMTS uses WCDMA.

What we are used to call UMTS is GSM + WCDMA (the standard can go farther, but the bread-and-butter is GSM + WCDMA now).

GSM uses TDMA (which is not the "old" TDMA, which technically is called IS-136, and IS-54 when it was part AMPS, part TDMA).

The WCDMA used by UMTS (when your phone is in "3G" or "3.5G") isn't your grandpa's CDMA either, which technically was CDMA2000, and IS-95 before it).

I specifically said WCDMA because it is a characteristic of WCDMA (and CDMA) to be much more sensitive to uplink power control than GSM when it comes to degrading overall network performance - when your UMTS phone in (W)CDMA mode transmits with more power than it needs to, that is "noise" for everyone else's phone signal in the same (W)CDMA carrier trying to reach the cellsite.

That effect is not as relevant when your UMTS phone is in GSM mode, because you are not sharing the same frequency (at the same time) with others near you. In (W)CDMA, many phones share the same frequency at the same time, each one being separated by a XORed code, crudely speaking (black magic involved).

If your phone in GSM transmits with more power than it should, you just get worse battery life (and a remote possibility of interfering with someone using anoter cellsite far away).

If your phone (or a lot of phones) in WCDMA transmits with more power than it should, it hurts network capacity.

That sounds strange. Only while in idle mode (no calls in place) a GSM/UMTS phone has some autonomy to select the cell site to which it will "listen" to.

If the iPhone follows the UMTS standards, while in a call it reports received signal information from neighboring cell sites (or towers) to the network, and the network then decides which cell site(s) the phone will use (or switch to, what is called handoff). The phone has to obey the network's decision, so the only way that the phone software could affect the cell site choices would be to send "fake" (or wrong) received signal information back to the network (which would violate the GSM/UMTS standard).

...so when the call quality gets bad, you'll be able to blame the network, not Apple.

...and on the transmit side, the software will pump up more average power than what the network power control requests, so your battery life will get worse, and overall network (WCDMA) performance will be degraded for all users in the neighborhood.

Also consider the possibility that this could get you in trouble in a math/physics exam. I thought I'd never go back to a classroom, but was pushed into a post-grad course - and was prohibited to use my old faithful HP49G on the financial/accounting exams "because it is alphanumeric and can be used for cheating"! I had to borrow a 30 year old 12c, but you won't be able to borrow a clean pair of arms.

Very interesting. Like the underlining rules that form different fractals, these political/social differences shape different maps.

There could then be at least three "drives" competing to shape the grid: "uniformity" (theoretical ideal mold global solution, unaware of traffic needs), "traffic shaped" (overall, prioritize the majority, screw the few who needs odd itineraries), "incremental" (each line is negotiated as an isolated problem/cost, overall final result may be unoptimized) - besides the obvious "economic/geographical" (how much it costs/is it possible to go through or around a mountain/lake).

I'd expect the same to happen in other systems where location and capacity planning are necessary: roads, schools/hospitals, telco backbones, cellular phone base stations, etc.

We are complex molds, but one could tell the kinds apart by looking at the drawings we make in the petri dish.

Somehow the two maps you mentioned fit the expected, one showing a more cooperative/harmonious society, and the other a more individualist one!