Yesterday Mozilla launched a new project, called Prism, on their Labs page. Prism allows web developers to build web applications that run as desktop apps - they get their own icon in the dock, they get their own menu, and they launch in a version of Firefox that has no browser chrome (forward, back buttons).

Initially this sounds a lot like AIR, although there are a few differences. First of all, Prism currently has no mechanism to work offline, though they are working on including it. Count on it using functionality from or similar to Google Gears, given the close relationship between Google and Mozilla. Secondly, it won't have access to read local data natively (at least it isn't mentioned explicitly). This is something I think is a major advantage to AIR - the ability to get to the local file system means you don't have to upload files to the server and then download them back to the client again - you can access them directly.

However, some similarities certainly exist, and its nice to see people acknowledging the same problems - its good validation for what AIR does. First of all, there's no browser chrome in either an AIR application or Prism, so the back button and bookmarks and other browser characteristics disappear. These don't make much sense in applications (though its interesting to note that the eBay AIR application actually added those elements in based on user feedback). Secondly, there's some benefit to being on the taskbar or dock, both to users and to websites. For users, its easier to launch the application, and that's also good for sites making those applications - if its easier to get to, people are more likely to use it.

Mike Chambers has a great blog post on the tone that the Mozilla people take with respect to Prism. I agree with him completely - it does sound like what Adobe is doing is evil, and what they're doing is great, even though AIR can do exactly what they're doing (and more). Sounds like we (Adobe) need to do some more education on the fact that AIR can run HTML / Ajax apps without having any Flash integration at all. In fact, there's nothing in AIR that requires you to write Flash or Flex code - you can write an AIR application in only HTML and JavaScript. (In fact, one of the first AIR applications was GMail on AIR.)

The advantage that Mozilla has for Prism is that they can integrate it into Firefox, and hit about 30% of the web community immediately. They're relationship with Google means that the technology will likely be implemented by Google applications, like GMail, which gives it immediate credibility and a ton of installations. The advantage to using Adobe AIR is that Adobe has a great track record of getting software installed on a lot of computers - the latest version of the Flash Player is installed on 98% of machines, and we can use that knowledge to get AIR installed on as many computers as possible, and as quickly as possible. While not the only consideration to be made, developers must consider how many people can use their applications without having to install additional software programs. I think currently that's an advantage for Adobe, but time will tell how well Mozilla does at getting Prism installed.

There's lots of good news here for developers though. Google Gears (if that's what Prism uses) and AIR use the same SQLite database, and work has been done to standardize the APIs. Adobe donated the Tamarin runtime to Mozilla, so coding should be similar for Flex AIR applications and Prism applications. Perhaps most importantly, there's choice in the market, and competition always keeps people motivated to work hard and improve things.