Yes, yes. All the cool kids are talking about Live Mesh. I'll get to that; this is related.
There was a very interesting post on TechCrunch a couple of weeks ago entitled "FriendFeed, The Centralized Me and Data Portability". It's really struck a chord with me.
It's introduced the concept (buzzword) of the "Centralised Me", which is lovely marketing, but might very well be a bit of a red herring.
The thinking goes like this: back in the day, all you had on the internet was your home page, and all your random thoughts, photos of your cats and interesting links went up there (usually edited by hand, in raw html. Hardcore). Nowadays, there's Flickr for your photos, YouTube for videos, Facebook for your friends, Twitter for your inane babble and so on. In other words, we've gone from having a very centralised view of "me", to a very decentralised view - "me" is spread across many sites.
The first problem I see with this is that it's not quite true. We might not have had Facebook, but we did have, for example, Usenet and mailing lists, both highly decentralised. Getting a single view of all of my interactions would have been a daunting task. And what about blog comments? Again, very decentralised.
And let's just think about that for a second.
Centralised and decentralised are just points of view.
All of my comments to blogs are very decentralised, but to the blog owners, those very same comments are completely centralised - attached to the blog posts they're in response to.
Facebook messages are centralised to all my friends on Facebook, as my Flickr photos are centralised to my Flickr friends. They're just decentralised to someone looking for "all" of my stuff.
Even if you look at the poster child of decentralised authentication that is OpenID, it's only decentralised as far as the protected web site is concerned. As far as I'm concerned, I always log in at the same point - my OpenID server. Centralised. All Yahoo IDs are now OpenIDs. Centralised.
So where does that leave us with social networks?
FriendFeed is an aggregator of your other social networks. You join up, and start broadcasting an aggregated view of every other social network you're a member of. You subscribe to other FriendFeed members, and you've now got a single port of call for all the updates you're interested in. It aims to be the Centralised Me.
But to quote the TechCrucnh article, this just means it's another "data silo".
And this brings us to the Data Portability Project. Ideally, it should help to protect us against data silos. In theory, as long as each silo implements the correct Microformats, and authentication (OpenID, OAuth) we should be able to access and copy/move our data out of a silo.
What I haven't seen is where we move the data to. Another silo?
The most interesting thing I have seen in this space is Google's Social Graph. It indexes the Microformat information found in web pages, and automatically builds up a social graph from this (see also Microformat's social network portability). Extrapolate a little here, and you can easily see how this technique could get a single view of my entire social network. It wouldn't matter where I put what, Google would find it for me. Google would find me. Not centralised, but aggregated. However, it's not as easy as that. Google can only access public information. It would be great for Twitter, and irrelevant for Facebook. We could give it credentials, but then it becomes another data silo.
There is no Centralised Me. There's just convenient and inconvenient.
Or rather, centralised and decentralised don't apply here. It's not a hub and spokes model. It's a (gulp) mesh.
Subscribe