Bit of discussion on Twitter today about whether the decentralized web is a pipe dream or a near-term possibility. My response to that is longer than a tweet, so I put it here.
Many things about the decentralized web are hard. IPFS, the torrent-like file system that makes servers irrelevant, is pretty geeky right now. Federated Wiki requires some initial guidance.
What can happen right now, however, is the that the data layer and the application layer of apps and web sites can be separated. This is the idea behind the Berners-Lee project SOLID, and is the idea I outlined a couple years ago.
This is a simple idea. When used to by MS Word, I’d pay money to Microsoft for it, but then I would point it to my private hard drive. Corporations did well, but the difference was that I maintained control and portability of my data. Microsoft did not own my documents. I didn’t have to store them on a Microsoft drive.
I liked that world.
A few days ago, Jim Groom was prodding me as to why I still host this blog on WordPress.com. And it’s true — I hate it serves people ads. I’d like to get off it.
But what I hate more is that half my old posts are completely messed up because of a hack back in 2011. Back then I was hosting it on BlueHost, I got behind on updates, I got hacked, and had to recover posts from an export I had made in 2009. For posts between 2009 and 2011 I actually had to reconstruct them from archive.org captures.
I know, I know — there’s auto-updates now, yadda yadda yadda, it will never happen again.
But the question I have to ask, again, is why I can’t have WordPress or Reclaim or whoever maintain my web application layer while paying for commodity storage somewhere which provides permissions based access to that data on a per application basis.
Such a scheme, among other things, would allow multiple of applications to plug into the same data unlocking the power the web used to have. Third party apps could use standard storage APIs instead of relying on the charity of Twitter or Google. They would also give me direct control of my data. And because data does not have to be ported, it would eliminate much of the current lock-in we find ourselves in. And tool providers can go back to selling us products instead of monetizing data. I think we’d all be happier, especially Jim Groom, who wouldn’t have to look at ads like the one below.