This article is absolutely ridiculous.  It acts like Google has done something to avoid the control of Apple.

This was Apple's original vision.  They wanted people to just build web apps for the iPhone.  They are overjoyed at this as it is proof of what they were saying from the beginning; it is possible to build a rich dynamic app using just safari on the iPhone.

The article then goes on to act like this web app will work on any mobile device, which is a patent lie.  Only webkit-based Android and the iPhone have the technology.

This is an interesting article about technology, platforms, applications, and content delivery.  But it is very short on any market assessment or critical analysis.

It is true that Apple's iPhone was initially released with web APIs and they supported web-based development from third parties.  In many instances (CNN, FoxNews, Wunderground, etc.) mobile platform were created by large content providers.  The same will probably be true for the Google OS as well.

However, native APIs that can take full advantage of device hardware have proven to be much more attractive to developer than web-based APIs.  Richer content can be created and manipulated by the user versus web-based APIs.  In addition, surfing the web is a cell phone convenience, and not superior to surfing on a computer which can provide a higher fidelity experience to the user.  It is a convenience.

When Apple released the SDK for the iPhone and simultaneously released a content deliver channel for the iPhone (iTunes App Store), they accommodated to the needs of the customer.  The customer wants a boutique-like shopping experience (as can be seen with the Amazon web site and iTunes music store), they want tight integration so that the experience is tight, seamless and fluid.  They also want an easy and standardized way to pay for content (as can be seen also at Amazon with 1-click and of Apple's online storefronts).  With an iPhone you can connect to the App Store, search for content, read reviews and see ratings, make a decision, and then quickly pay for the content and have it delivered to your phone by only typing your iTunes account password.  This is a very convenient user experience.

It is much more difficult to log onto a web site with a mobile device, enter your credit card number, personal billing information, and click through a few windows to pay for content.  Count the amount of clicks and characters you have to type to purchase software online versus purchasing it from an iPhone.  There are dramatically fewer steps on the iPhone, and that makes it easier for customers to purchase products.  Again, it's about the user experience.  This is where the Google OS diverges.  The OS works great but once you leave the OS, the user experience can very greatly.

In addition, the iTunes App store is a marketing channel.  You can put up any web page and embed a link to the iTunes store which brings the user to the product where they can purchase it quickly.  The store also contains external links to developer's web sites.  It's a two-way marketing channel, and one of those channels is provided by Apple.  Support and maintenance of the channel, which can reach millions of users, costs only 30% of an application's sales.  In many cases developers spend in excess of 30% the cost of the product to market it without a guarantee that it will sell.  30% is a large margin for marketing, but the iTunes App store provides more than just marketing.  It's an interactive feedback loop for developers.

I also do not think that Google's platform will be as successful as it's touted to be.  We have yet to see significant Linux desktop adoption and that was touted to dethrone Microsoft.  Linux does well in server environments or other embedded applications, but it does not deliver a seamless user experience.  Google will experience the same symptoms because it is allowing developers more freedom to create unique user experiences.  Like in the early days of Windows and more commonly in Linux, the lack of consistence and control creates user fatigue and they become disinterested in the platform.  Handheld manufacturers will also fragment the market further by writing custom code for their handsets that deviate away from any standard-setting Google may try to impose at a later date.

When you allow multiple parties to do whatever they want with something, they will all try to make it unique from the next guy.  In some cases it will be spectacular.  In other cases it will be an abysmal mess.  In all cases, it will be just okay (think about a bell curve distribution).

It is for the reasons listed above that Google's phone OS many not be as successful as you've described above.