Diaspora Project: Building the Anti-Facebook
readwriteweb.com
readwriteweb.com
I don't think software people understand just how disinclined a non-software person is to do anything requiring configuring serverside stuff.
Even torrents, where clients have essentially become good enough that all you have to do is install them, most people get a friend that has some understanding of computers to set them up for them.
The cloud is really a better model if you want to achieve widespread adoption, and what is the point of a social network without people.
The largest cell phone market is the one for generic simple phones that just make calls, regardless of the smartphone hype. The reason that IE is so ubiquitous is that a huge number of people don't really care, as long as they can get on the internet, and go on facebook and youtube and maybe a few other sites, thats enough for them.
Hell, even look at dropbox, people only started using that sort of service when it went into the cloud and became dead easy. They didn't care that they were locking all their data into a hard to leave service.
Don't get me wrong, I think it would be great if there was a way to recreate facebook in an open software setting, I just don't think any social network where the average user would have to either install a complicated app or pay a fee will ever get off the ground...
I see people not caring as a lack of education. Computers are ubiquitous in the western world since less than a generation. The Internet is even younger. This is just too short for people to understand them (and their hazards). Just wait 20 years, and they will understand. At this point, they will care. Now, the question is whether we can accelerate the process or not.
even if those non-technical people don't host their own data, they'll be able to switch to something else that may have different privacy policies.
They still surf the net, send email and talk to people on the phone though.
For instance, in a fully decentralized system, you can get rid of storage limitations by storing stuff on your computer (or on the cloud in a private form). So you can share not only pics of your pets, but also your music and movies collection.
If this results in something as easy to use as Facebook, but uncontrollable and perfectly able to host and protect media piracy / political dissidence / etc., we get darknets for dummies. This project or a variant thereof, if successful, could deeply change Internet by taking all power out of authorities and into people's hand. Whether it's a good or a bad thing is debatable, but it certainly is interesting.
Not, not really. It's a very good thing, and similar to what the web did in the first place.
Sorry for the sarcasm & general negativity - but they have nothing. As someone else said, just build it. The concept of replacing Facebook with something open, private and distributed is not at all new - many have tried and failed before them. "We have the chops" and $10k isn't some kind of magic potion. And their biggest challenge isn't building it, it's getting people to leave Facebook for it.
I won't be the first to call it now - but I can imagine the popularity of decentralized architectures catching mainstream attention. Git is getting popular, vs say SVN or CVS.
Remember, the internet itself is decentralized.
Also the ability to easily move your identity and inform people you have moved so they can update. Doing that last part securely is one of the things that hung me up when thinking about it, but I also only really dug into it for an hour or so.
I think DNS is the key here; it lets you move a server while keeping the same name.
Of course I may be stuck on the way I was looking at it of a fully qualified URL (say http://runevault.com/runevault for my id, and I don't think DNS currently goes farther then domains).
A standard such as this would be crawled on a frequent basis by the social network apps built on top of it, as well as the search engines. They would all update their canonical profile addresses and that would be that--an organic, self repairing network.
Getting people to pay for an otherwise free service might be tricky, but I'm sure google would foot the bill (and do the hosting) if they got to push more ad-words.
It will probably go: self-hosted and hosted (by techies), then big companies find a way to do the hosting for non-techies in order to push ads.
Movable identities - DNS works. A DHT with email-addresses as the primary key (rather than domain names) would also work (though you would need some karma ratings system to keep it from being hacked). You would either use DNS, or reinvent a DNS-like protocol. No big deal.
I think it's a common misconception. You can be decentralised with friends. With the whole world, you need a single entity. Just look at Facebook, Twitter, Google, MSN -vs- Jabber (even gtalk is not a jabber in the mind of users - it's gtalk / google chat), money transfer services, iAnything from Apple, etc.
People seem to like centralisation - they know where to go and how to call it.
Email and web? Not so much really. Everyone serves their little page from their own site. But as soon as you want to interact / share, you're back to hubs like dropbox, facebook, twitter, etc. The data may flow through decentralised links, but it's still coming from FB.
If I want to send you an email, I talk to my server, my server talks to your server, and you get an email in your inbox. Maybe I use GMail and you use Yahoo, but the underlying protocol is the same, so it doesn't matter.
Services like Facebook and Twitter have been written as a closed protocol right from the start, so you're not used to thinking of them in open or interoperable terms, but it looks like it's happening - it's just a question of time and the right impetus.
(Perhaps you're thinking of 'decentralised' as being able to host stuff on any computer in some sort of cloud, but I prefer to think of it as 'not controlled by any one entity' - maybe that's where some of the confusion is coming from?)
My plan is similar. RESTful API built on OAuth 2.0 and Open Social, with OpenID as an option. The specification will be open and the first implementation will be OSS.
The idea is that there needs to be a protocol such that I can friend you, even though you're hosted on a completely different server. I suspect that it's a difficult concept to wrap your head around unless there's something that you can actually play with (although it's not too dissimilar to email). Perhaps the Django/Python choice might throw people too...
Even for a decentralized project you still need the big "seeders", that is the problem.
(Though I don't really get marcus's point/joke, probably because I haven't watched the video.)
Also, by re-federating the service of Facebook with partner companies (push pics back to flickr, status back to twitter, etc.), they could get some nice allies, reduce the bandwidth requirements dramatically, and allow granular control of users in a very clear and powerful way (flickr is set to private, for instance).
Neat idea. I'd love to see it kill facebook. If there's an elegant way to detach online socialization from the profit motive while maintaining functionality, it would be good for everyone not employed by facebook to see it succeed.
Really, it seems like turn-key services are the primary solution, though the important thing is to get a competitive marketplace where some are ad-supported, some aren't, but all of them are reasonably trusted.
That said, would it be possible to get research universities to act as the network? I assume people would be trusting of that.
Or is there a way to encrypt data or split it up in ways that makes it meaningless to any individual without the right password/signature?
There are ways to combat this. Using OAuth for authentication keeps them from getting your credentials. It or a similar protocol for data exchange allows you to shut them off if they're proven bogus. They might have limited visibility into your network in the meantime, but that's it.
It will probably end up working a lot like the web works now. You'll have to look at the URL and decide if you trust the provider. You can also send a message along with the friend invite and prior to accepting another's invite, that gives you some confidence that you know that person. Finally, by seeing that you share common friends on common providers will be a strong indicator.
Diaspora server will be the same. Your random 419 scammer won't be able to make a lot of friends if people are a little careful. Add PGP keys and a proper web of trust, and the scammer won't be able to spoof any identity at all.
Ultimately, a larger subset of your data will eventually become public. This will happen by users installing games and other apps on top of the system that aren't as trustworthy. This can happen today with Facebook. It just isn't talked about much. Writing an app for FB is eye opening.
After we open source our source code, we hope to also provide a paid turnkey hosted service in the vein of Wordpress.com to make it easy for people who want to use Diaspora, but don’t want to deal with the fuss of setting it up.
Apparently it is ok to take your users for a ride, but only so far.
I'm pretty happy about all this, maybe there will be a serious push back against all these companies that see the information they hold about their users as their 'product'.
What's next, distributed search engines?
Interesting idea, how might that work?
Also, it was funny that youtube included onesocialweb.org in the related videos at the end - it seems to be a similar concept, but I've never heard about it before.
If they are hoping to be able to hook in to existing semi-closed systems like Facebook, though, they are aiming for a moving target that, as we've seen, isn't afraid to use the law to stop users from getting their data out.
I also wonder how they will handle communities. That is one thing that livejournal has always done quite well, and facebook's groups leaves a lot to be desired. I have a hard time imagining how that will work in a distributed environment, unless the community itself will not actually be distributed.
I'm sure that it's not just tech geeks who would like to have some modest amount of privacy online. Just declaring that "Privacy is dead. Get over it" may be ok for a few exhibitionists and philosophers, but in the real world people do want and expect some level of privacy - especially when it comes to things like medical conditions or relationships.
Will anyone ever install this? I doubt it. Good thing is the people who actually think this is enough of a problem to install some server side stuff have the technical capability to do so.
I almost just typed in: Solution in search of a problem. After reading comments here + around the web I feel there is some semi problem here that might apply to a broader market- Facebook is getting creepy with how much they know and it's getting overly complicated. I just think this is a really poor way to solve the problem.
You Create your own data and keep it stored on your device(can be anything, laptop,desktop,mobile etc). When you visit some host, it knows where to find your data(we assign a common known to all social port to our host), but needs your permission to download the data.So now when a relation of yours visits the same host we visited, It can tell our relation that what we have done on the host, only after we both use a common key.
Yes, there are many social network style applications waiting to be discovered, but this effort, Diaspora, not only has a lot of attention, but it hit at a good time - when people are openly dissatisfied with the incumbent service, and the team heading up this project has not only the support of groupies, but their school, open source software organizations, and a lot of very motivated community members.
Lets hope that we can build something that will endure for the long term, and reverse the trend of user exploitation.
Anyone know if that effort is still under way? (and if there are any lessons to be learned from it?)
kind of funny that's an option
from my understanding, the name implies that people who don't have a social networking account from some service provider are expatriates. i think this gives too much importance to such services.
nice idea, i definitely would like to see how it goes. but the name should be changed, so that it is in line with the political view it is the brainchild of.
I think a home page will play a part by integrating into the service. Instead of having a Facebook home page, you might go to any of a number of sites that would provide an aggregated view into your portion of the social fabric. Instead of looking at peoples profile pages, you would go to their profile page on a blog or anywhere else. This, too, would integrate into the underlying service.
This is all supposition, of course...
That aside, it looks like they're building in GPG based sharing (to limit _who_ you're actually sharing with—a more and more common problem these days) and some kind of capability to automatically publish to/subscribe from friends' nodes.
It doesn't seem to know about graphs of friends, which is the central principle of FB. FB is not about videos, YouTube is; FB is about getting the videos you like to spread across your friendship network.
However, it looks like a nice application, even if it's only remotely related to Diaspora or Facebook.