James Governor over at Red Monk (a great bunch of analysts I alternately seem to agree with and take issue with) posted a take on the mobile platform development continuum and HTML5 as an alternative. Take a read and the disagreement I voice below will have a better context. I’m reposting the reply here because it is a relevant discussion and I leave him the right to do with his site as he will, and that includes having it go away. So I’ll keep my work where I can get at it. On my laptop in MarsEdit and on the web on my own blog as well. 🙂
[I feel] James, [is] getting a bit absolutist for an analyst that usually sees how things fit together more completely. It’s not a winner-take-all end game. Some apps work better on a local platform, native toolchain, optimized into the hardware, fixed system, no browser barrier. Usually have a higher development cost than the average web app. These are, of course, generalizations.
Just look at the Samsung Galaxy response. Some people like it. Different form factor. Different purpose. iPad app onto a Galaxy? Most wouldn’t feel right or be a great experience. The web gives you a “common denominator” approach, and that usually gets offset by a pile of javascript conditionals adjusting for the nuances of browsers.
I tell you, trying to use a mutual fund screening system on the web today was horribly painful. The state and ability to pull things in and out of a spreadsheet app is just not happening. The app is older and could be improved with some newer approached and technology, but really, I want my <em>data</em> to be able to flow back and forth securely between devices, but when I’m working, I want it to be <em>here</em>. That’s why the browsers keep adding in desktop features. Because the experience in the browser is and has been inferior.
WebGL? Offline Storage? Hardware graphics acceleration? Codec hardware acceleration? These are all desktop features that get pulled into the browser over time as it trails behind. It’s not surprising, and it’s not going to change. Standards move slower than most proprietary innovation. Web apps are usually for a cheaper and/or broader approach, with a radically different revenue model.
It just baffles me that development gets lumped into a bucket of “proprietary platform” or “web standards”. Both are two ends of a continuum. OpenGL is pretty cross-platform in the bulk of the API. And it’s well defined on experience. There’s shades of gray all over. Are you telling me the same technology and approach for Farmville is appropriate for Doom3 or Call of Duty Black Ops? I don’t hear the game companies complaining about supporting 5 or 6 platforms, let alone 2 or 3.
Of course, the desktop pulled in automatic updates and notfications and sync from the web in many ways to bring those strengths across too, so it’s not totally a one-way street.
But it’s not a one-lane street either.