I hate to say it but I would never use an iphone web app over a native one. It's less about looks than it is about feel / responsiveness. You also will always be stuck dealing with the forward button / back button structure.
You are not stuck dealing with the forward/back button structure. iPhone web apps can be bookmarked like applications. The author can tell the iPhone that it wants to be launched in full-screen mode, without the nav bar at the bottom of the screen. (I believe this also works on Android)
Furthermore, web apps can be created that explicitly manage data and assets in a local store. Web apps can be quite responsive if they take advantage of this.
A native app will likely be more responsive, but a web app isn't as limited as you seem to
think.
I think there's a need for a high end cross platform toolkit/framework for developing web apps that provide some high-end stuff for iphone and android but also can degrade gracefully for "phone" phones that have crappier browsers. Ideally this would work like unobtrusive javascript, instead of having to have two views.
I'm currently investigating building a better mobile web presence for an application. The solution for us so far has been to mangle WURFL to a faster format and use that to detect properties of the current device from the UserAgent (width/height mainly). The WURFL way is to do a lehvenstein calculation on the current user agent versus all the agents in the db, pretty expensive computation.
This puts the crappiest phones first, because the web we build is very restricted by the poor phones, but the truth is that iPhone/Android is rapidly approaching 50%, because they are more likely to actually use the mobile internet.
To build this could quite possibly be a javascript layer on top of a basic REST structure.
With the success of the Appstore, Apple also has little incentive to invest heavily in enhancing the web interface. In the long-term, web apps will likely be better supported on Android, since the web browser is Google's major focus.
1. It's possible to write your app within the constraints of a Web App
2. You have hosting that can handle the potential load (how many million potential users?)
3. You plan on doing something that is likely to get rejected by Apple
4. You plan to give the app away for free
I think allot of people overlook #2. For me one of the nicest things about writing native apps is the fact that I can focus on coding and not infrastructure, and avoid the risk of having too much (wasting money) or too little (can't service customer spikes) bandwidth, etc.
The demo of the jQuery-based jQTouch Beta shows how an iPhone web app could have a visual look similar to that of a native app.
Yes, you'll never recreate Tweetie 2 with it, and it provides a lot of eye candy. The animation demos (page flip) are impressive. Web apps aren't as sexy as native apps, and they can allow you to "run" on multiple mobile devices, with pretty much the same server-side code.
Yeah, jQTouch is actually a really viable alternative to making a native app. It makes use of the native animations and features for a really polished interface.
Because the market is much, much smaller. I don't know what percentage of iPhone users actually jailbreak their phones, but looking at the iPhone users I know, I can only think of one person that jailbreaks. It might not be worth the effort of actually making a different version just for people who jailbreak. If app developers really did embrace this, that might drive more people to jailbreak, but I highly doubt your average Joe going through the process of jailbreaking, even if it isn't particularly hard.
On top of that, there might be fears of piracy. I don't know how valid it would be, but targeting users who jailbreak is also targeting the only people that pirate since you have to have a jailbroken phone to pirate.
"...targeting users who jailbreak is also targeting the only people that pirate since you have to have a jailbroken phone to pirate."
Bingo. You'll end up with your Cydia-enhanced version sitting next to your pirated retail version side-by-side on the Cydia store. If you want to target "pirates" you can't use traditional sales methods - they are a totally different channel.
I would shudder at the thought. You already have it built and have no use for the iphone development environment so there is no use to accept the hefty split in purchase price with Apple. Only positive is getting listed in the App Store. And the usefulness of that is getting more doubtful as the number of apps increases.
