Much has quickly been written about the just launched Amazon HTML5 "Cloud Reader." It's a "web app" that could be used to replace the native Kindle iPad and iPhone apps. It's being widely read as a response to Apple's more restrictive App Store terms about in-app purchases/subscriptions. (Amazon has removed the button linking to the Kindle store from the most recent version of its iOS apps.)
I agree with John Gruber and others who believe this has been in the works for longer than Apple's more restrictive terms around in-app purchases. Cloud Reader does, however, completely side-step Apple and its App Store terms. The Financial Times and Walmart's Vudu also recently did something similar. More publishers and developers will follow, who don't want to give money to Apple or simply don't want to worry about Apple's rules.
An added benefit is compatibility with all mobile platforms. However developing native apps is relatively easy these days because, unless/until Windows Phones break out or BlackBerry stops its slide, only two operating systems matter: iOS and Android.
In contrast to some of the rapturous reviews I've seen Cloud Reader is not as strong as the Kindle native app. It works well and allows users to read books offline (e.g., on the iPad). But it's not as responsive or fast the native app.
Yet for certain types of sites or publications HTML5 will be fine and quite usable. However for higher functionality a native app is better, even necessary (e.g., games).
But now that several companies have "validated" the HTML5 strategy will we see the migration from apps to HTML5 that the industry has always anticipated? Not exactly. I think what we'll see is continued development of native iOS, Android (later maybe Windows Phone) apps and HTML5 from everyone else. However Apple doesn't help its cause with all its rules and restrictions.