I’m not adverse to having a standard presentation layer, but I am concerned about the increasing girth of HTML, CSS, and Javascript (see my post to @david). Most wireless carriers are planning to implement quality of service controls based on policy in the next 2-3 years. What that translates to is degraded performance for large consumers. I have yet to see anything in the current web standards that indicates things will become more compact. They just keep getting bigger, and that will become a real problem when those controls are in place.

But your mileage may vary.

Granted, I’m not trying to write an online office suite or photo editor, and no doubt there’s a lot of browser-specific JS in Google’s code that runs when I embed Google Maps, but the browser-incompatibility problem on the desktop is 99% solved for 99% of sites. I’m hoping we’ll be able to say the same for browsers on mobile devices very soon.

BTW, if you design your site properly, JavaScript and CSS will cache, so the extra bandwidth shouldn’t be an issue (maybe a few KB/month/app).

In a similar vein to your post, I wrote a post last week suggesting three things that browser developers could do that would bring us much closer to HTML5 Apps being a reality: http://blog.romeda.org/2010/05/three-simple-things-that-browser.html

I remember, a bit over a decade ago, when the same thing was true on the desktop. Desktop browser compatibility is still far from perfect there (IE6 is the monster that will never die), but in general, write-once-run-anywhere is a reasonable goal for a desktop web app, and it’s easy to do a web site that runs well in IE7+, Firefox, Safari, Chrome, etc.

With both iPhone and Android using web kit, and RIM about to move to it, it might not take nearly so long to get the same level of compatibility for mobile sites — non-web-kit browsers are going to be forced to follow the same standards and specs as web kit or risk being left behind, and there are many more accepted standards in place now than there were in the late 1990s. It’s already very easy to create a mobile site that runs nicely on iPhone and Android, unless, of course, you expect it to look like a native app instead of a mobile web site.

Web apps are good for some things, native apps are good for others. Solutions like PhoneGap, Appcelerator, etc. provide some degree of “portability” with regards to creating native apps, but if you dig down into the guts, you still have to potentially customize the UI for each device, which should come as no surprise at all. There are performance issues as well that are tied to cellular network performance (which is almost always worse than WiFi or a LAN connection), processor speeds, memory, etc. Also there are the issues associated with browser implementations (and don’t even get me started on the fragmented nature of JME on devices).

Pundits claiming that web apps are the future are speaking from a desktop/laptop perspective. They obviously don’t get the issues around web apps on mobile devices. It’s a nice goal, but it hasn’t happened yet and I’m not sure that it will happen in the way the rosy picture has been painted.