To app or to not app | Hacker News Reader

To app or to not app

benrameshlin.posterous.com

25 pointsrameshnid19 comments

Open articleView on HN

Things the mobile web is currently missing:

* Native performance.(getting better, still not there).

* Offline is way harder (possible, with manifests, but harder).

* Monetization is harder: you can't just 'sell' a webapp.

* The appstores (can be) a great distribution and promotion channel.

* The web has way less access to the device: No contacts, camera, etc.

It's getting better, but native apps do have their advantages.

Wait. Why do you have to choose?

What's wrong with the following pattern: 1. build an HTML5 app 2. build a native app that embeds a Web control 3. package the app and HTML5 app together using application cache 4. distribute on AppStore

You get the a) offline support b) monetization c) distribution via AppStore

the html5 app won't be capable of using device specific input, such as accelerometer, compass, multitouch, etc. All of that is not yet part of the html 5 spec. So when we compare html5 vs mobile apps it's just an apples against bananas type of comparison

Well, the strategy i'm describing involves a native app. In that case, would it be possible to expose device-specific API to JavaScript?

Also, multi-touch is actually supported by webkit. There a nice WWDC video that intoduces this capability if you'd like to learn more.

To be clear the idea above isn't new of course. For instance http://www.appcelerator.com/ provides a native app with a thin wrapper around a web control.

Phonegap provides a thin wrapper around a webview. Titanium changed course with their 1.0 version, and now compiles the Javascript to native code. (they used to do the webview thing).

2. Browser apps when powerful will compete with the apps on the app store.

ie, right now they don't. I don't plan to wait.

Patience pays, most of the time.

If I was making a product that wouldn't be done for a year, I'd think about targeting next gen tech. But for apps that take at most a month or two? No way.

I tried making an iPhone web app... tooo slooow.

My limited experience so far echoes this. I'm in the middle of scratching an itch my wife has in such a way that it's useful to more people than just her, and part of that is a mobile application.

Sencha (and jQTouch before it) is absolutely gorgeous; I wanted very much to fall in love with it. On a webkit browser on my desk, it produced visually stunning results.

And then I tried using it with both an iPhone 3G and a G1. Yes, they're old tech, but they're the hardware I have on hand right now to test on (and not exactly uncommon in the field), and performance is abysmal.

I would assume the same tests run against an iPhone 4 or an HTC Evo would blow me away, but my wife has a G1, and I'm building this for her first. :)

visit www.gmbhnews.com from your iphone/android/BB to see how it looks as a webapp in a field where you mostly find native apps. It may be a bit difficult to market mobile webapps from my experience, since peoples don't use search on their smartphones (I think, this stateent is from Steve Jobs), but it's definetly worth it. and you can still create a native app , if it's really the point, and target a specific platform.

Is there any realistic way to do a mass-market GPS-driven app outside of an app store?

Hmm. Don't know if there is an option completely outside of an app store. U can use the app store that Tom Tom plans to enable on it's new gps device.

See http://www.electronista.com/articles/10/04/29/tomtom.to.use....

What percentage of the market carries Tom Tom devices? It sounds like a tiny, tiny percentage. What would I be doing there other than proving the market for someone with the stomach to make the real money on the iTunes store?

True. Or u can kill ur tom-tom market with ur own iTunes app. The thing about tom-tom is it's a gps device. So u get easy adoption for a gps service. Because you get the pre-established intent to use a gps service.

Basically get noticed with the Tom-Tom(more likely because of pre-established intent) and then come up with an app for other devices. Or like I suggest a browser based app( I don't know if this is currently possible, but am sure it will happen soon)

visit Postabon.com from Safari on an iPhone

If all you need is coordinates, there's always W3C's geolocation API: http://dev.w3.org/geo/api/spec-source.html

I don't think so, but you can still try to keep as much logic as possible on the web. (E.g. by providing a browser window to your website and a button that does a POST to your API, and nothing else...)

Postabon's mobile website can access the users GPS location, if that capability is present on the user's device. smanek can chime in here. (I wrote their iPhone app.)

The javascript location API works since iPhone OS 3 and Android OS 2.

Reply on news.ycombinator.com