"Native" really isn't that significant. When Android first came out, it's "native" applications were all apps that were written more or less from scratch using the Android APIs. When Android first came out, there were essentially zero apps. "Native" usually isn't raw ARM binaries either, it's in the form of dalvik vm bytecode, optimized not for speed, but for executable size. The only time native is useful is with truly performance critical applications, and the vast majority of apps aren't performance critical. Javascript runtimes are getting faster with each release and the huge amount of competition between browser vendors. Chrome also has Native Client, which would be the equivalent of Android's NDK.

If by "native" you mean the user interface consistency, acknowledging that web apps will be significant pretty much destroys that anyway. Lots of iOS/Android apps don't even use the native UI widgets anyway, so it's not a really worthy goal.

Android started with a huge disadvantage: not having any apps. Yet only a few years later, the platform hundreds of thousands. Chrome and the web as a platform is much more approachable, and for every existing android or iOS app, in a short time, there will be rough equivalents that can be made.

From the start, Chrome has millions. All of them can be accessed with a consistent interface, where there's a single notion of an application. If android takes over, then you'll be constantly switching contexts from the browser and the "native" applications.

Have you spent much time with an Android phone? Android's strength is how seamlessly you can switch contexts. It's easy, it's fun, and it's efficient.

There's definitely a place for ChromeOS, but I'd rather use Android, because the browser just plain sucks at some things.

This rather assumes that the browser could not be improved upon.

I kind of feel like Tom... Chrome OS is actually what Android will turn into, not the other way around. The downloadable app paradigm is working fine for Apple, but it's been pretty miserable for Google. Their marketplace is pretty atrocious, and the method of selling them has been a constant headache.

I think Chrome OS is a stopgap to an OS where the browser doesn't appear as a bunch of tabs, but as something that probably looks more like iOS anyway. I think it shows Google thinking a little too small and releasing too early. Had it baked a couple of years, and it became branded as Android 4 or something ("Android... everywhere"), I think that would have been better. I wonder if they're not launching early in order to push things like Web Print, so by the time they are ready to do something groundbreaking, the technology is already in place.

"If Android is brought to the point where it supports web apps built for ChromeOS there will be very little reason to develop native Android apps."

From what I've seen, the gap between a native Android app and web apps is getting bigger, not smaller. Among other things, the Android API includes: 1. A SIP-based VoIP protocol stack 2. Near Field Communications 3. Gyroscope and other sensors 4. Multiple cameras support 5. Speech recognition engines

Even if these things were available tomorrow to web apps, developers might still choose the native option because a) There would be millions of 'old' devices that only supported the native API; and b) There would be thousands of developers who were familiar with the native API

There's accelerometer support in Firefox, Chrome, Opera, and Safari (at least mobile Opera and Safari) now. It's only a matter of time until we have useable APIs for NFC, camera, etc. I'm wondering when there's going to be a browser USB API. I can think of some things I'd like to do with that.

Native apps still "use the cloud". The only substantial difference is the language and APIs used to develop them.

But they don't have to use the cloud. Meaning native apps can (and in my experience do) leave data on the device which in turn means the user experience isn't seamless from device to device. You're so focused on the technical aspect you're missing the philosophical side of this which is the important part. By forcing app developers to use the cloud you create a guaranteed seamlessness and that's the whole point.

Web apps don't "have to use the cloud" either. (there are APIs for local storage)

I don't dispute that the "web apps only" approach his more "pure". My prediction is simply that the Android approach will win out because users value functionality over purity.

Part of the reason web apps aren't saving much data locally is because they don't have anything to save (where's that API for accessing the camera?). BTW, the fix for native apps saving locally is to make all of the local storage automatically sync into the cloud, much like dropbox does (and like iPhone does with all of the app databases when it syncs). I think that's how the Sidekick worked (which was also created by Andy Rubin).

To address your points...

Web Apps don't have to use the cloud - I suppose it's possible for a web developer to decide to locally cache all an application's data but it seems like your grasping at straws here. It would require a developer to specifically disregard the overall philosophy of a web based apps and a Cloud OS. Don't get me wrong. In my perfect world Google would force developers not to do this but as it is I don't think it's something you're likely to see much of (my understanding is that Google is focusing on SD usage for media just to prevent this)

On users value of functionality - Again the question comes back to one of "what functionality will they really be losing". Particularly with WebGL and Flash both available to developers I really don't see a functionality gap.

Forcing Android to the Cloud - If you just desperately want to save Android you could absolutely do this but what's the point. If you're going to force everyone to the cloud and you can build your development environment around web based tools than why wouldn't you? More to the point why would Google spend a bunch of time and money turning Android into a Cloud based OS when they already have a Cloud based OS

I'm not trying to "save Android". I'm simply predicting that Google will continue developing and improving it because they are already shipping millions of Android devices and it has immense strategic value to them.

I suspect ChromeOS syncs the nominal 'local storage' up to the cloud, as well. (Or if it doesn't, it will.)

The API for the camera is right here :) https://mozillalabs.com/rainbow/