Adobe AIR: Do we really need browser apps that run on the desktop?
In his post, he describes his personal user expectations:
I generally am not that inclined to download and install the applications…He notes Buzzword as an example:
I love RIAs, I love the ease in which I can access, use and share them…
I just don’t think running an RIA as a completely separate entity to the browser works too well. A large draw for an RIA is simplicity and availability, both of which out-of-browser RIAs restrict.
Take Buzzword, which is a truly remarkable RIA and in my opinion is superior to any other online AND offline word processing tool. I’ve not seen the AIR version of it which I believe is on its way and I can already see that what AIR offers will be a great addition to it. But wouldn’t it be even better if all the connectivity detection, local file storage, drag and drop etc were available to the in-browser version?Danny-T does see a need for the features that are available in AIR, but not available to Flash/Flex in the browser:
Okay, so the features that AIR offers aren’t available to the Flash player and I do think those features are hugely valuable. If installing AIR gave me all those capabilities as an in-browser resource as well as a stand-alone runtime. I know it’s not so easy with all the security sandbox constraints of running in the browser.Adobe’s Ryan Stewart responds in his blog:
Security gurus can tell me that what I’m asking for isn’t possible and why it shouldn’t be, but I feel as long as I ultimately make the decision on what does and doesn’t have access then it shouldn’t matter whether this is in or out of the browser. Am I alone in thinking this?
I don't totally disagree with him. As cool as AIR is, I still think the browser will be the central point of contact for most web applications… but even with some of AIR's features, the applications in the browser won't be desktop applications, and that's where I see the major benefit of AIR…David Coletta from the Buzzword team notes that the chal lenges go beyond security on Danny-T’s blog:
To me, AIR represents a true hybrid approach to application development. It uses web technologies and development trends but allows developers to create real desktop applications. That actually may not be valuable for everyone...
The key question is what do your users want. Do they want an application that lives on their hard drive, installs like any other application and behaves like any other desktop application should (offline, with the file system, notifications, custom chrome, etc) or do they want their application in the browser?
Sadly, it’s not as simple as just working out security issues. There are many technical issues with browser/Flash interaction that prevent Buzzword from being all it can be within the browser. For example, rich text clipboard and keyboard management is practically impossible to get right inside the browser, though Buzzword comes darn close.For those of you in the InfoQ.com community, are your users asking for features and functionality that will make the desktop integration and offline features of AIR valuable? If so, do you need all of AIR, or is the offline support of a browser plug-in like Google Gears enough?
For more information on Adobe AIR: http://www.infoq.com/air
What the fuss about installing?
Perhaps the question is: Do we need a new view on what an OS is? AIR and Firefox 3, as well Java become OSes. There is a lot of effort made to work around the simplistic "extension model" of Windows &Co which has the open all or nothing attitude (ActiveX sadly promotes this into the webworld)
I can't deny that my favorite distribution model is Java WebStart - not perfect (ie transparent) but the truest
Re: What the fuss about installing?
I think if you're going to go with that model, you need to dig deep and consider building some things out from scratch, enhancing others and building tools to help you get over the rough spots.
I do hope that the new emphasis on consumer/client helps to smooth out some of the bumps.