Gizmodo has reported on an experimental user interface from Sony called DataTiles. It's physicality is cute but seems impractical. However, the level of integration is impressive, and obviously (to me) has a lot of potential.
I think it's very hard to integrate all of these different bits of functionality, for two reasons.
First of all, different groups of people have to agree to use a common standard / architecture / protocol in order for new functions to emerge. This gets worse when these people are from different industries - say, IT, banking, commerce, telephony.
Second, it is not yet quite clear what these new functions that could emerge are, and how people will use them. Naturally, the latter increases the difficulty of the former.
An open, bottom-up, build-it-and-they-will-come approach seems to work well - after all, this is how the Internet was built. Web services, the current incarnation of that trend, are a good example. Google started their Google Maps API, and already there's a ton of ideas about what could be done with it, and with mapping in general. Even games!
Now some of this is moving from the browser to the desktop, with Google Desktop, Apple's Spotlight and Dashboard, and, at some point, Microsoft's online and offline search technology.
Is the browser better for experimentation? The OS makers have means, motive and opportunity to integrate the functionality of their systems. Apple, which has more control over their products and its context than Microsoft, also achieves better integration - one of the reasons why I'm a Mac user. Both companies have created an ecosystem of independent software developers and the APIs and tools they feed on. And yet...
Search is getting a lot of press right now, but searching is only one aspect of integration. Growl is a unified user messaging system for Mac OS X applications. It's slowly growing, but it doesn't seem to add a huge amount of value (I have it installed). Still, it shows the kind of architectural decoupling that is required to integrate disparate bits of technology. Applications don't show events to the user: they report the event to something else which shows it for them. Much like an API in an OS.
I find Dashboard the most exciting project along these lines. The makers claim Microsoft stole their concept and is calling it 'implicit query' - I bet they're real happy about Apple using their name too (for an OS feature which looks remarkably like Konfabulator - oh well). Dashboard shows a vision of where Spotlight and Longhorn may end up, and it's fairly compelling. (Or is it? I've used Spotlight once since Tiger came out, and I could've done that search with Panther as well.) It's going to be interesting to see how all of this develops.
I've just realised that this blog entry has drifted to a subject only vaguely connected to games. Games tend to create closed-off universes in order to increase immersion, both physically (with consoles) and in terms of software (taking over your PC's screen). But still, connections exist.