The two main benefits I see of the app store model:
- Centralised program repository. (Users can go to one site and find programs listed in a consistent format, with related services provided alongside, like ratings and reviews.)
- Program "quality control". (Spam, malware, and unusable programs aren't admitted: developers must make their programs good.)
Googling for programs, then having to find out how good they are on Wikipedia, forums, and blogs is a pain. The app store model means users can easily find good programs.
(Admittedly the app store model has some flaws but I'm playing devil's advocate here.)
As for the app model, making every little function its own program can seem overkill, but this allows them to be tailored to a device and have greater consistency with the device's look and feel.
The app store gold rush is certainly cause for cynicism, but I do find myself missing it on other platforms. A known point of discovery; a community rating system; and streamlined purchasing is convenient, as is not engaging the cognitive overhead to figure out how to buy something, pay for it, worrying about credit card details, registering for new sites, etc. Maybe this is Amazon's future, instead of a separate one per platform.
I'm not talking about the distribution method. An online marketplace for digital products is not what defines the app model. The app model is the idea of small individual programs for each individual task/need/amusement/whatever, with minimal interoperability or extensibility.
That is a matter of definition. Another way to look at that "small individual programs" and "minimal interoperability" are just a historical side-effect of iPhone (hardware, software, and licensing) limitations.
With iPad, nobody forbids you from selling an office suite as one tool. In fact, Apple's Pages for iPad may be an early example of this, with its limited charting capabilities.
Looking ahead, I would hope that we do not go that way, but that, instead, evolution will be towards a component store, where editors are sold, and viewers are free downloads (backed by the app store, something like Opendoc on iPad could be a killer)
What would it mean for the app model to be dead? The app model is basically the outsourcing of deployment and billing to a service provider. It's not a specific technology. You can do that with web apps just as well and I don't think this idea will ever be dead.
The question is whether this model is as suitable for business apps as it is for consumer apps. I think the IT department is not dead, because someone has to integrate all the various apps and services and analyse the growing quantity of data in ways that are specific to each company. These functions are too heterogenous to be cast into yet another app.
Hardware provider => platform provider => application provider. This is the hierarchy of the cloud-era Internet. Think AWS => Engine Yard Cloud => Zendesk.
Hardware is becoming a utility. AWS is billed exactly like electricity: pay for what you use, use it however you see fit. Platforms follow this model too, with a healthy markup.
Apps are what really matter, and what drive value from a platform and hardware. Money comes from customers paying for apps.
Maybe my comment was a bit ambiguous. These days 'App' can cover a wide variety of concepts. I'm referring to the 'App' focus of things like the iPhone and the iPod (bite size, one-off programs), not 'Web Apps' (which are more like services).
Until Mac OS X and Microsoft Windows require using an App Store with only approved software available for "licensing" (the new term for "install") then most of the computer market will continue with business as usual.
I bloody well hope not.