First, I don't know why you guys focused on the NFC part of the article I linked, when I was obviously referring to the other part
I do expect Apple to come up with something like Google Wallet (using iTunes) to bypass carriers there but that's not what I was focusing on and it's orthogonal to how you handle the SIM (it simply requires an additional on-board secure element chip).
Arun,
You had some comments earlier about the GPS choice in the iPhone 4 and iPhone 4S wrt respective accuracy.
Can you elaborate on that? According to the (excellent) Anandtech review, the 4S has a better GPS. Is this a surprise to you?
I know next to nothing about the subtleties of different GPS technologies and implementation, so your insight would be appreciated.
With all due respect to Anand, the rest of his review might be excellent, but his GPS tests are really stupid. He's mostly driving in a straight line in an area without high-rise buildings. I'm not sure what he's expecting to happen - even a 5 year old mid-range GPS chip could handle that situation fairly well.
What Anand is actually testing is 'consistency' and that's based on both the quality of the RF and the associated software. While a better RF would indeed be valuable, it's very typical for the GPS chip itself to 'smooth' the signal (i.e. the OS/processor never see the raw data), so it's impossible to tell how good the RF is based on so little information. You might as well judge a 12MP camera with a 3MP sensor based on the quality of its interpolation algorithm.
Now in practice I do expect the RF to be better in many cases because of the GLONASS support which I frankly forgot about but maybe Broadcom is still superior in other ways, hard to say. Certainly based on user reports the iPhone 4 CDMA was not as good as the iPhone 4 GSM despite what Anand claimed, but I suspect based on Qualcomm PR release timing that it did NOT support GLONASS (at least not with iOS 4, maybe it could be added via software). There's also various sources of interference that will affect RF quality but the iPhone 4 & 4S designs are so similar I doubt it matters here.
The main problem is things like 'urban canyons' where you need a lot more than good RF and signal smoothing to get a sensible result. Being able to access some extra Russian satellites can't hurt but it won't fix the fundamental problems. You need some pretty clever (and not free) algorithms to make it work AFAIK and my guess is Qualcomm is still lacking on that front. Basically my point is the iPhone 4S might be extremely confused in the middle of a large urban center (with an extreme case being something like New York City). I'd certainly be very curious about a test in such an environment.
I'm not saying the GPS is bad though - FWIW, my mother is buying one and going to use it as a PND replacement with the TomTom app at my suggestion. It's probably 'good enough' for most people. Just don't take Anand's test as the end-all be-all of GPS performance.
---
Talking of GPS - one of the best smartphones for GPS today is probably the Samsung Galaxy SII since it uses the SiRFStarIV. Coincidentally, CSR also announced the SiRFStarV today - should give you a pretty good idea of the future of location technology in general:
http://www.csr.com/news/pr/release/682/en
Summary: supports GPS/Galileo/GLONASS/Compass satellites simultaneously, automatic anonymous tracking of detected WiFi hotspot locations on a CSR-owned server, 3G/4G assistance, Accelerometer/Gyroscope/Compass MEMS support for indoor positioning, combining all of the above. Let's see if they actually deliver but needless to say it's about a lot more than just raw RF accuracy nowadays - that's basically a solved problem to be honest.
Oh and on an unrelated technical note: the SiRFPrimaII uses a standalone RF chip (unlike SiRFStarV) but that chip is made on... 40nm! So much for CMOS RF not scaling fast enough to be cost effective
