Where are the killer RIA apps?

"Killer apps" was a term originally coined to describe apps that drove hardware purchases - 123 for the PC, Pagemaker for the Mac. It seems to me we live in an era of ubiquitous hardware, so perhaps the concept is obsolete.

There is a fundamental problem with RIAs; the open web is a much better platform to build applications on. Every day browsers get better and the things you can do with HTML, JavaScript and CSS are for 90% of the applications a good solution. The other 10% of the applications a developer has a choice where RIA might make sense but 5% of those applications are probably better off being a desktop applications.

This leaves 5% of the applications which would benefit from an RIA, these apps are typically internal data entry or dashboard applications which don't get a lot of press.

Over the last 6 years I have worked in the RIA space and it really comes down to a simple question. "Would you want Google to rewrite the Google Analytics or Gmail in Flex/JavaFX/Silverlight?"

For me the answer is NO!. The best applications today are open web with Ajax / Flash enhancements to improve users experience.

Google Analytics uses Flash and Ajax in this exact way; Flash for the charts and the time span widgets but the rest of the application is HTML. Ajax is used for doing sorting and paging of data.

This is why there will never be a killer RIA application.

Typically RIAs take longer to build and polish than the slightly more lightweight HTML/AJAX combos (which some have argued are killer apps too.) If you want to create something big and you see relatively simple web apps dominating with RIAs often becoming bottom feeders (twitter clients are a great example) then what technology would you choose to build the next big idea?

Many RIAs do not run as smoothly as they could because they are jam packed with graphics and effects or there is just too much information flowing between client and server to run smoothly. This does not allow you to churn out many quick prototypes to try and hook users. Also, for many users standard internet applications are rich enough. I have seen plenty of really cool RIAs but none with a chance to play int he mainstream (save for maybe the Twitter clients.)

Will there be killer RIA apps? The only way I can see it happening is

1) when they become light-weight enough that they become preferred apps on light-weight devices, i.e. cell phones, pdas, MIDs, etc. As right now, they're more processor intensive (at least Silverlight and Flash/Flex/Air are) than apps written with more established technology. Causing battery life to be dismal.

2) Internet connection becomes truly widespread. (Only spelunking caves and Montana should be the places where you can't get internet connection.) And not only widespread, but fast, too. No point in having Rich Internet Applications if the Internet part is a part of the bottleneck.

Their are great RIA, but it's the app that is more important. I started using Prezi yesterday, and it is truly amazing. But it isn't amazing because of RIA, it's a killer because of the product, and it just happens to be a RIA (isn't RIE more suitable? Rich Internet Expeirence).

It's like asking 'where are the killer python scripts'. the technology doesn't matter as much as the product. 123 for PC, and Pagemaker for Mac where killer apps not because of the platform they were created on , but because they replaced the previous methods of accomplishing the task.

Patience. Patience. I'm working as fast as I can.