Android will not unseat iPhone
I'm a big fan of my iPhone. Being by nature somebody who wants to know everything immediately all the time, a device that lets me Google every 20th question that occurs to me* would be very dear to me no matter how it worked, but the iPhone is extraordinarily good as a web tablet, and it works fine when I need a phone**.
As is the case whenever you're a big fan of something, human nature is such that people will good-naturedly try to poke fun at it or otherwise deflate you. So I hear a lot of people -- especially Googlers -- telling me that my precious iPhone is soon going to be dead, because Google is going to turn up and dominate the mobile market with the awesome, open, Android operating system.
I'm going to make a bold prediction: Android will not gain greater than market share than iPhone. Possibly never, but definitely not within the next 5 years.
This is not a conservative bet. Keep in mind that iPhone beat Windows mobile, a comparable system that had been in the market for nearly 8 years, in market share within 6 months of launch (at least in the US), and achieved half the market share of the 11-year-old, business-focussed Blackberry, despite being a consumer-focussed device. If Android is better than iPhone, it should be able to get significant traction within 2 years. But it's not going to.
Closed vs. Open
The competition between Android and iPhone has been billed as a competition between an open operating system vs. a closed, single-vendor system. The assumption is that, being open, Android will simply have so many more eyes on the problem that Apple will soon be eclipsed. The model for this assumption is Windows OS versus Mac OS: Microsoft lets anybody build hardware and software for their platform, leading to massive competition and better prices. Apple strictly controls hardware and is pretty picky about who builds software too, and -- the thinking goes -- is thus less popular, because it is more expensive and less versatile.
The variable that is not being factored in here is hardware. A market where OS X is making great progress is notebooks, where it is now the number four vendor, growing significantly faster than its overall market share (although it's number 3 in that market, mainly because the market is so much more concentrated by Dell and HP).
Why is Apple doing better in notebooks than desktops? It's because while desktop macs are just high-end desktops, notebook macs are better laptops for the same price. And the reason is because Apple controls the whole stack: on a laptop, where space and weight are at a premium, owning every piece of the manufacturing process means you can integrate things more tightly, shaving a gram here, a millimeter there.
In the mobile space this effect is multiplied. Phones are like computers were 20 years ago: starved for RAM and processor, and pretty bandwidth-limited too. Here Apple's ability to combine everything really tightly is absolutely killer, and it shows in the competitive size and weight, and market-beating performance of the iPhone. It is also the key feature of the usability of the iPhone: each app can rely on the existence of the other apps, so Safari knows how to play video via the "YouTube" player (really a generic video player) and music via the iPod, and the SMS app can rely on Safari to exist when you click links. It all comes together.
Android has none of these advantages. It has to be flexible to run on hundreds of possible hardware configurations, and modular to support all the possible features of all these devices. It has to include security and management down to the lowest levels. These all add precious overhead, both to the size of the operating system on the phone and the amount of memory it takes up when running, and that reduces performance. The lack of a single overarching design team will complicate the user experience and produce a less pleasing, consistent UI. The modular nature of the system will introduce confusing configuration options. These are not things you need to try out Android to know. This is the fundamental nature of an open system: they work best when you have resources to spare.
What will Android's role be?
Android will launch, and geeks will love the extra features on the HTC Dream, and crow endlessly about all the things they can do on their Google Phone that you can't do on iPhone, and it won't matter. iPhone will have superior performance and a better user experience, consumers will continue to prefer it, and Android will be stuck in a niche. As phones increase in power and capability, the advantage of an integrated system will diminish, but not as quickly as you think: Moore's law may make processors faster all the time, but it doesn't apply to battery life, already the principal limiting factor in mobile hardware.
If Android goes anywhere, it will be into the role currently occupied by Symbian, the operating system that runs Nokia phones. Many have predicted a merger of the two systems, but that seems very counter to Nokia's strong build-your-own culture. Nokia instead seems to be looking to fight Android on its own terms: it bought out Symbian from its partners to gain control and promptly announced that it was going to open-source it and give it away, just like Android. It's going to have a tough job, since Symbian is awash with performance-sapping legacy features, while Android is new and shiny, but it has a gigantic incumbent advantage since most phones are literally built for Symbian from the ground up.
But if Android took over Symbian, you say, that would be huge! Symbian has 65% market share! Sure, but that's going to start dropping as Apple continues to expand iPhone into other countries and inevitably diversifies the range into a higher and lower-end model. Android will be competing for a slice of a shrinking pie against fierce competition from a deeply entrenched and extremely strong competitor. The flip isn't going to happen in 5 years, and it might not happen at all.
So I reiterate: iPhone will continue to grow because the closed model works better, and Android will either grow very slowly or die on the vine. You heard it here first (maybe).
* If I Googled every question I would be Googling 3 or 4 times per minute, and I simply can't type that fast.
** Call records indicate I make or receive a total of 2-3 calls per day, seldom longer than a minute. I recognize this is atypical.