When I first heard about Google's and then Mozilla's concepts of "installable web apps" I was a little dubious. Surely the whole benefit of web applications is that they don't need to be installed!
But if you think of installable web apps as a convergence of the "app store" and "cloud computing" models (arguably the two most significant trends in the software industry today) it actually makes quite a lot of sense.
The diversity of different software platforms available for smartphones, tablets, smart TVs and PCs is great for encouraging competition and driving innovation – but it also causes problems for both app developers and consumers. An app written for Apple's iOS for example will not run on Android, Blackberry OS, WebOS, Meego, Bada, Symbian or any other competing platform. For consumers this is confusing and frustrating because unless they allow themselves to be locked in to a single vendor they may have to re-purchase different apps for different devices. For app developers it's even more frustrating because they're faced with a bewildering array of platforms to target their software at.
Web applications have different problems with limited discoverability and access to hardware. To a web browser web applications are no different to any other web site which means that the apps can easily get lost in the ocean of content on the web. Web apps also traditionally have very limited default access to client-side hardware for security reasons which limits their capabilities when compared to native applications.
Installable web apps can combine the best parts of web applications (the benefits of cloud computing and accessiblity from any device) with the best parts of the app store model (easy distribution and discovery of software which has full access to the device's hardware). This new breed of apps could leverage web technologies to reach the whole range of competing platforms with a single application which can be distributed via a range of competing app stores. By "installing" a web application the user can pre-approve access to local hardware like offline storage, accelerated graphics and geolocation to allow the app to make the most of the hardware it's running on and the user can keep a local repository of all the applications they use.
Both developers and consumers only have to worry about one kind of application running on one open platform (the web), but still have a choice of hardware and app stores. In this model the only people who lose out are the corporations who seek to lock consumers into their own products and services.
It could take a long time for installable web apps to become the predominant model of software distribution, but the idea of converging cloud computing with the app store model makes the idea very appealing and might go some way to explaining why Google currently has both the Android and Chrome OS operating systems in development. Each OS is approaching the convergence from a different direction whilst maximising on the market potential of what can already be done today.
I hope that standards for "installable web apps" can be agreed upon sooner rather than later so that if this model does become a growing trend, the new breed of apps will not suffer from the same fragmentation as other platforms do today.