[javajosh]

I was with you until you said "This platform cult is pretty meaningless to me." Actually the platform wars are quite meaningful insofar as they represent large-scale manipulation of consumers. Both Apple and Google are playing the following game: hey consumer, here's a shiny device, it's pretty cheap and then we lock you into a 2-year contract, meanwhile you acquire apps that can only be run an iPhone, and now even if you wanted to go to Android, you are faced with the prospect of ditching your entire software library (or vice-versa). It's called lock-in and it's a long-term manipulation of a market. I'm sure that both Apple and Google rationalize that it's a way to recoup R&D costs, and that's true enough, but it's a sneaky way to do it.

There's also the not insignificant problem of app security. Not so long ago there was a news item about the FB native app accessing contacts and uploading them to FB servers. The web app cannot do that, and that's a Good Thing.

There's also another reason why HTML5 apps are better than native: they are composable. There isn't much out there other than search engines that takes advantage of this fact, but webapps can be composed, parsed, and otherwise uniformly manipulated in useful and interesting ways. Android and iOS do indeed offer some integration points for their native apps, but nowhere near at the level that the web allows today.

[blub]

The web app already has all my data on a server somewhere. And that web app is 99% of the time closed source. And the company can cut me off, change the TOS or sell my info to advertisers. Now THAT's a lockin.

While you identified the problem, the solution is a truly open mobile OS, not webapps.

[javajosh]

How does a truly open mobile OS (and applications, presumably) solve the problem of vendor-specific data lock-in? Data lock-in is a totally separate, and far more difficult, problem. Part of the reason is that our data a) probably wouldn't exist, and b) probably doesn't make sense out of the context of being mixed together with everyone else's data. There is a strong a priori centralizing force for some important kinds of data.