thanks for the reference to Dion Almaer's work.
it's interesting to me because not long ago, i tossed together an experimental mozilla prism bookmarklet to generate desktop webapps on--the-fly (http://prismspectrum.com/bookmarklet).
i do believe adobe air and similar projects are indeed great contributions to the RIA scene... that is obvious. but it's also refreshing to see advancing concepts in hybrid apps and app helpers like Mozilla Prism or other variations that make use of XUL/XULRunner. Prism got my full attention when it was announced because it aims to do things that have always interested me.... from simple chrome wrappers to file uploads and data synchronization. And although building a companion application using Adobe tools is a powerful approach, as a developer, I would first take the path of building my web application so that it also can work with a Prism or Google Gears wrapper (or both even).
I hope to gain some time to experiment further with Prism and XUL. I'll also be providing Mozilla Labs (and anyone else) with the bookmarklet code that i wrote (with mozilla preferred licensing) soon.
fun stuff.
Sull
Dion Almaer has a post over on his blog about what some of the future Google Gears APIs might look like. This one specifically is about creating desktop shortcuts. It’s important to note that this is very, very early so this API is indicative of where Google Gears is going it isn’t considered anything close to a final product or news item. Dion surmises that a common use case for Prism and AIR is “just a wrapper around a Web application that you use all the time”. This Google Gears API would mimic the behavior of Mozilla’s Prism and let you do a single click install which would provide a desktop shortcut. The way Dion describes it you could put an “install” button anywhere on your site and then invoke the APIs to create a shortcut on the desktop. You could give it an icon and possibly even tell it to open with or without browser chrome. Viola, a desktop application….sort of.
It’s clear that moving to the desktop is more than “offline” mode. While that made for a quaint argument building real desktop applications with web technologies has become compelling. Prism is following that model and it looks like Google Gears may be moving in that direction as well. I think all of those moves are good because they provide a better web experience. But I think there are some fundamental issues about what role the browser plays in situations like this.
The browser clearly has a lot of traction. And it makes things simple. You get a basically instant connection to any application out there and as it expands in functionality it starts to take away more responsibilities from the underlying operating system. It isn’t inconceivable that someday browsers could support file extensions for individual sites so that when you click on a file it opens that web site and passes a reference to the file so the browser application can use it. And that’s great, but as a web user do you want your browser becoming bloated with operating system functionality?
Maybe you do, but I think in the zeal to eliminate the very compelling niche of the web enabled desktops the browsers are getting distracted. Web browsers are great for lots of things but instead of focusing on adding desktop functionality it would be great if they focused on standards. It seems to be working for the IE8 team. Of course Google Gears isn’t a browser, it’s an addon. It’s feature set, bringing browser applications offline, is great. But as the footprint expands what happens to the browser? Will it still be the tried and true tool that we all love?
I wouldn’t be surprised if one day the browser really is the gateway to everything but until that time the desktop will remain very important. It’s how most people access applications and it’s filled with tried and true user interface metaphors that everyone has grown up with. As a developer I’d encourage you to take advantage of that. Adobe AIR puts that into the hands of web developers which is why I was sold on it so early.
