Hyperbole

There has been a lot of hullaballoo in the Mac developer community about Apple's decision to make Safari, or "Web 2.0" <cough> the "SDK" for the iPhone (in lieu of an actual "SDK", maybe even some sort of XCode plugin.) Without going into the miry mud of this debate, the burning question I have about this revolves around the part of the iPhone technology I would most like to experiment with: will the iPhone Safari browser make it possible to write a JavaScript app/widget that can respond to multi-touch gestures.

As I see there, there are three ways they could expose this:

1. A high-level API with a collection of hooks for the various gestures used in native iPhone apps.
   
   (e.g., pinch horizontal/vertical, spread horizontal/vertical, chord horizontal/vertical, and maybe some idioms for new single-touch patterns like toss up/down, or whatever they're calling it.)


2. A low-level API that gives raw (or barely normalized) data about gestures
   
   (e.g., reports position pairs)


3. Both of the above

As I see it, if they could expose both high and low level tracking for gestures, they would gain immediate benefit from developers implementing rich, iPhone-specific UIs for their web apps. They would also gain from the having the developer community be able to invent new gestures, most of which will likely be confusing, but some of which may actually be adopted by the iPhone in future software iterations.

Granted, I recognize that writing code to recognize new gestures based on the low-level API I describe above is no small matter. Consider something as simple as the "chord" effect being complicated by the diagonal chording gesture (currently supported in MacBook(Pro) trackpads): you would have to disambiguate finger fluctuations in such chording from the motion of spreading or pinching either by timed modes, tolerance thresholds, or some combination of the two. Not for the faint of heart!

Anyhow, that's what I was wondering...

Still haven't decided if I'm in for version 1 (err, mass-market public beta 1, more likely.) Some more details from AT&T would be helpful.