Hacker News Reader

Comment by cookiecaper | Hacker News Reader

I've been thinking about doing some kind of distributed/open social network thing what with all the Facebook scandals cropping up, but it's pretty hard to do.

The fact is for something like Facebook to function smoothly, it has to be centralized. There has to be a server somewhere out there that can be queried reliably for this data and that's the bottom line, even if you have something API-based like OpenSocial. It's just a matter of who you're going to trust with that data -- and as we see over and over again, money and power clouds judgment; a good candidate to control this network when it's starting out may not hold up so well under the barrages of billions in revenue and investors down the neck.

The only way for users to have real control over their data is to build some P2P/self-hosting architecture, and for reasons all HN readers understand, this would be a huge, huge mess.

View on HN

The fact is for something like Facebook to function smoothly, it has to be centralized.

I hate to do this to you, but [citation needed].

for reasons all HN readers understand, [P2P] would be a huge, huge mess.

Count me as one HN reader who does not understand that. (But then, it's hard to get someone to understand something if his ideology depends on not understanding it...)

I've explained a few obvious problems with a decentralized, P2P architecture here: http://news.ycombinator.com/item?id=1330382

At first glance it seems nontrivial. As with most problems like this that can be decomposed into algorithms and datastructures, technology provides a solution. OpenID, OAuth, and Open Social were all developed to solve similar problems.

My guess is that there will be half a dozen competing protocols within a year.

Better to have one good one.

Right, and how successful are those protocols you've named? They all have relatively limited deployment and usage even though they've been around for years. OpenID is a nice idea but people don't want to remember a URL and a username and a password. People don't want to remember a "master URL" to give to each new social app they sign up for -- they want what Facebook gives; just sign up with a traditional username and password and start finding your friends immediately.

And, again, even then, if we follow the OpenID model of external providers, it's all about trusting that external provider with your data. There's still a central point with your data on it. There has to be. Unless the user is self-hosting this data, there is always a reliance on external entities. And how long could an external provider survive merely as a host for applications like Facebook, while not pulling in any revenue from Facebook's revenue channels like ads or Credits?

What would be a mess about it? Couldn't we be clever enough to figure out how to do this with what we have?

Reply on news.ycombinator.com