HTML5 is a lovely platform for cross-device development. Basically, it's the only game going forward. But it's really not an answer for building a great app for a given platform. Apple is
talking up HTML5 in order to
combat Flash, but it's just talking about web sites, not apps. HTML5 will rule the [moribund] desktop, but for mobile devices I think it has major challenges.
HTML5 does not get the same level of access to the device that you need to build a rich experience.
- Integration with the contact list (is there anything really more important?)
- Access to phone status, history and actions
- Camera(s), proximity sensor(s), microphone(s), accelerometer, compass, gyroscope, multi-touch, speakers, etc (and a lot more to come)
- Local storage, access to SD-card files, application backup and restore
- Native configuration and management interfaces (sync, preferences, phone migration, privacy, network gsm-vs-wifi, etc)
- ... drumroll please: the app stores. This is the channel for getting apps for these devices. Otherwise they have to find your website somehow.
HTML5 apps will be good enough in some cases for all devices, but they'll always be step-children to the native environment. You could argue that we just don't care about the weird sensors and whatever else HTML5 doesn't give us access to. I disagree: the really useful apps for mobile platforms will take full advantage of these features, recording and correlating all sorts of information and drawing conclusions from it - where you are, where your customers are, when you're together, what you're carrying with you, how your spouses are getting along, what you've sold them, transcripts of your conversations, your body temperature, what they've bought recently, voice stress analysis, who else is around, what's mouldering in their warehouse, and what expression is on your face. Science fictional? Sure. "The future is already here, it's just not evenly distributed." Mobile technology is going to level the playing field for these kinds of intelligent applications.
HTML5 will continue to evolve and will slowly add access to mobile functionality common to all devices, in a lowest-common-denominator way. The fact that
Webkit will be on Blackberry by the end of the year makes HTML5 a cross-browser contender - it will lock up the entire mobile landscape, making cross-platform browser apps even easier. But so far, geolocation is one of the few things that work cross platform. The full list of things above will come over Steve Jobs' dead body. [I'm only half-joking.]
PhoneGap is the only cross-platform development environment that currently has any viability at all, and it's risky because Apple routinely rejects PhoneGap-based apps; although they're written in JavaScript which is technically allowed, Apple takes a
dim view of anything not
*originally* written in Objective C. I don't expect Apple to be changing direction and opening up their platform and their store. If RIM survives [fat chance] its app store may go in a different direction - but I'm not holding my breath: RIM is completely beholden to [evil] carriers.
The app stores are large and getting crowded. But the publishers, labels, studios and carriers are in bed with the Google and Apple markets, and they have real legs. The markets are evolving extremely fast, especially Google's (which is moving into music and movies, and even has meta-markets like
AppBrain) and Apple is
s-l-o-w-l-y migrating to a non-desktop iTunes store. The smartphone market is exploding, and every one of these devices has an icon on the front screen for the app store. I don't expect these stores to go away any time this decade - there's just too much money to be made.
