The phone went on sale yesterday. Developers still have no way to write anything for it, and the official position from Moscone three weeks ago is that they do not need one.
The line from the keynote was that Web 2.0 and AJAX give you an application that looks and behaves like the real thing, runs inside Safari, needs no SDK, and cannot destabilise the device. Jobs called it a sweet solution. The room did not think it was sweet. I have rarely heard a developer audience go that flat that fast, and these are people who turn up predisposed to applaud.
I understand the reasoning and I think it collapses inside a year. The security argument is real: this is a phone on a carrier network, one bad application can degrade a cell, and AT&T will have written terms about that which nobody outside the deal has read. Shipping a locked platform for version one is defensible. Telling developers it is the permanent answer is not.
What kills it is not the complaining. It is that a web application cannot reach any of the parts that make this device worth writing for. No camera. No accelerometer. No local storage worth the name, no addressbook, nothing that works when the signal drops, which on EDGE is going to be often enough to matter. The entire argument for this phone over a Nokia is the hardware underneath the glass, and the sanctioned development model is the one model that cannot touch it. You end up with bookmarks that render nicely.
Somebody will break it open regardless. That is not a prediction requiring courage, it is a Unix machine in your pocket with a filesystem and no sanctioned way in, and the people who take that as a personal challenge are already working. When they succeed, Apple gets to choose between fighting its most enthusiastic users indefinitely or admitting the web app position was a shipping decision dressed as a philosophy.
Leopard got the rest of the keynote and I have less to say about it, partly because a delayed operating system demoed at length is not interesting and partly because everyone in that hall knew what they had come to hear about. Safari on Windows was the genuine surprise and I still cannot decide whether it is a serious play for browser share or a way of getting the iPhone’s rendering engine onto developers’ desks so they can build those web applications. Probably both, and the second reason is the one that pays.
My guess is a native SDK inside twelve months, announced as though it were the plan all along. I would rather be told now.