This isn’t necessarily a new idea, but it helps to have a visual metaphor to get past the API-geekspeak and look at where we’re really going.
I know that OmniDrive (“The Universal Web Storage Platform”) is already planning on integrating with Flickr as a “storage device”… so what happens when you literally can hook up to remove stores of data, media and so on, that aren’t restricted to FTP and SAMBA and so on, but typical web APIs? What happens when you can access event data on Upcoming.org not using an XML API, but by simply consuming the microformatted XHTML pages?
Essentially, any webpage or website becomes a data store and an application, all rolled into one.
It’s curious that when I first started creating this graphic, I wanted to illustrate the idea of networked harddrives — but that get their data from web services. Looking at it now, if each name had “.app” on the end, these would be a collection of desktop executables. So, with the line blurring further between representing applications and data, we start to glimpse just what a desktop-enabled web service access interface might actually look like…
Factory Joe 
Great post. Now if only we could all these web services to play nice together, and have some sort of centralized location for files, identity, etc…
Excellent visual for a complex idea; this would be a great interface…ok, not the Finder, but you get the idea 🙂
I can pretty easily imagine how this might be useful for something like OmniDrive, Flickr, WordPress. These are personal data focused web apps. But things like Facebook/MySpace are not so much data focused? They’re more about establishing connections and communicating.
Btw as for Flickr, one of my friends wrote this nice app for Linux called flickrfs which essentially acts as a virtual filesystem for Flickr. You can drag n drop your photos into your folder and it will automatically upload/sync them with your Flickr account. (http://manishrjain.googlepages.com/flickrfs). It is written in Python but depends on the FUSE package so I’m not sure if it can run in OS X.
@Harish: for Facebook/MySpace and anything else with social networking involved (i.e. connecting to other people) part of the contents of those harddrives would be access to your friends’ data… their photos, their blog posts, whatever…! The point of a socially networked harddrive is to give you local-like access in a familiar interface so that you manipulate the data you would like anything else on your desktop. And heck, if you can already browse your friends’ photos in a web browser, why can’t you do it in the Finder or Windows Explorer?
It’s a vision that’s becoming ever more desireable. However interoperability is not something that actually needs to be pervasive. For most people working in their walled gardens is just fine. Only when they move does portability become a concern. I do envisage a future where data flows between pools much more freely, nonetheless every application is still required to understand independant datatypes—unless abstraction becomes more transparent.
I haven’t investigated beyond their site, but Move My Data have a goal of adding a synchronisation/conversion interface to disparate webservices handling personal data. Much like I’m (slowly) doing with PictureSync, albeit I’m doing so on a semi-commercial basis. (As I think I mentioned I’ll release the MediaSock Client Framework—that will expose some of the interoperability of PictureSync with services—under a free-use licence at some point.) I’m off to check out the Revver API.
Testing