Slashdot Mirror
We Need Distributed Social Networks More Than Ello
At the end of September, "FacebookDragQueenGate" fell from the sky like a gift from the gods to the founders (and venture capital backers) of the Ello social network. The company promised not only to remain ad-free and to allow drag queen stage names, but even stated that they planned to allow pornographic content (something that received relatively little press, compared to the ad-free model). But critics such as Aral Balkan wrote that once Ello received venture capital funding, the backers would inevitably pressure the company to change its relationship with its users in order to make money. In an interview published in Forbes on Monday, Harvard Business School professor John Deighton was blunt: "The board will need to monetize the membership in whatever fashion ensures a profitable return of capital for the venture fund’s investors. So my advice, if they believe Ello is still viable by then, is to buy out [Paul Budnitz, the idealistic founder who came up with the 'no ads' idea]."
There is, in short, nothing to stop Ello from doing what Facebook does whenever they make a significant change to their Terms of Service: presenting users with a dialog box next time they sign in, saying, "These are the new rules, by checking this box, you are agreeing to abide by the new contract which you're not going to read." If Ello succeeds beyond its founders' dreams, then its ad-free nature might start to hinge on its founders all turning down buyout offers of tens of millions of dollars to stick to their ideals -- hardly a sure thing. Or the VCs might get enough seats on the board that they can outvote the founders and render their objections moot.
As Joshua Kopstein writes in an editorial for Al-Jazeera America, what really would have changed the game would have been a distributed, decentralized social network. I already wrote two pieces arguing that a distributed social network could work, and how -- a protocol that allows users to create profiles, "status" posts, groups, events, and other familiar social networking features as "objects" that live on their own server, but that can interact with users' profiles hosted on other servers. I don't want to re-hash all the details here, but the short version is that there seems to be nothing about social networks, as we currently use them, which would require all of the data to be stored in a single centralized system. In a distributed protocol, you could host your profile with any hosting company, and users could "subscribe" to updates from your profile, as well as the ability to receive invites to your events and your groups, and direct messages from you. Think RSS feeds, but with better support for well-defined objects like "event invites".
If your profile were linked to a domain name that you own, then if your existing hosting company ever deleted your profile (or threatened to), you could simply move your profile to a new hosting company, the same way that any person or company can currently switch their domain name between hosting providers. This, obviously, would instantly render moot any one company's policies about "real names" (or porn, for that matter) -- all you have to do is find at least one company, anywhere in the world, whose policies are permissive enough to host your profile, and that should be possible for all but the most extreme or illegal content.
This also renders moot all the worries about profile hosting companies trying to amass tens of millions of users and then stabbing them in the back, by changing the terms of service to allow them to sell user data or stuff unwieldy ads down their throat. When users can switch seamlessly between hosts, no one host is going to be able to "charge" more than the going market rate for hosting a profile (where "charging" could be in the form of monetary payment or displaying ads to the user). How much would it actually cost to host a profile for the typical user these days, complete with all their photos and status updates? It's hard to know, because other than university professors, nobody really has personal webpages any more, after they all went to MySpace and then to Facebook. But since the old days when people did actually host their own personal pages, hosting and serving data has gotten really, really cheap. For the average user, with a few hundred photos and a few hundred friends looking at them, $1 per year might be enough. Maybe they'd just have to watch one of those ads once a year that Youtube puts in front of a Beyoncé music video, and that would cover it.
Unfortunately, to many people the concept of distributed social networking is linked with the failure of Diaspora, the most ambitious attempt to create a decentralized protocol to compete with the likes of Facebook. But Diaspora didn't fail because the idea lacked merit; it almost certainly failed because people asked the same question that they asked of any other upstart Facebook competitor: Why should I join, when all of my friends are on Facebook instead? Of course people might reasonably asked the same question about Google+, but when Google launches a product, people join because they know the quality will be decent, they know that probably some of their friends will join because of the Google brand, and they know people will be buzzing about it anyway so they want to join in order to see what the big deal is.
And that brings up the story's second moral: Despite what you may have heard from your cousin who just read The Fountainhead, the products that are the most successful are not necessarily the best, by any objective measure; rather, they're usually the ones that had major backing (Google+) or were the beneficiaries of a staggering lucky break (Ello). Diaspora didn't take off, because it didn't have either one of these.
And since you cannot manufacture a lucky break, I continue to believe that the last best hope for truly free social networking -- with minimal censorship, and ads and costs kept to a minimum by market competition -- would be for a major player like Google to launch a social networking protocol, and to set up themselves as the default host for new profiles, but allowing the protocol to interoperate seamlessly with profiles hosted elsewhere. Either that, or if the system is launched by a startup or a nonprofit, make sure that you have a host of widely respected luminaries or organizations standing ready to help promote it -- if the EFF and the BoingBoing guys endorsed a new social networking system as the future of Internet freedom, people would join because it would seem uncool not to. As long as the product itself is functional, just have the right connections lined up when you launch it. Because that's what matters, and don't let the deluded ghost of Ayn Rand tell you otherwise.
Since when is it a requirement to sell ads in order to make a profit? Since when is "selling ads" the only way to make a profit? The entire premise is idiotic, because it presumes that "selling ads" is the only way to achieve cash flow.
If you create a service, and price it reasonably, you can charge a subscription / membership fee, and have a perfectly profitable business.
I pay for services all the time, why should an online service be any different?
Diaspora failed partly because it presents itself in such a confusing way. See Join Diaspora.: "JoinDiaspora.com Registrations are closed But don't worry! There are lots of other pods you can register at. You can also choose to set up your own pod if you'd like. There's no "Join" button, but two "Donate" buttons. Take a look at a few "pods". You can't see anything without signing up, and many sound like they're run by wierdos.
The latter is the real problem. A system where anyone can join anonymously and can have as many identities as they want will be overrun by spammers and jerks. Facebook has some pushback in that area, which helps. Facebook also started by getting people from big-name schools, so they didn't start with a loser-heavy population.
A social network needs some cost to creating an identity. The cost can be money, or reputation, or even a proof of work, like Bitcoin. Otherwise, the network is overrun with fake accounts. A distributed social network needs good anti-forgery mechanisms, to prevent one node from spoofing another. That's hard without central control.
Maybe we can pay Bennett directly to move his blog somewhere else.
Or maybe we can pay timmy to stop posting Bennett's posts.
Or maybe we can pay someone to write a greasemonkey script to hide Bennett's posts.
Money makes the world go 'round.
Agreed.
As I read this article, I was reminded of the push back in the 90's to get off the corporate networks (Compuserve, AOL, etc.) where data and people were walled off from opposing networks, and dive into the World Wide Web. At some point the pendulum swung back towards the value brought by corporate networks, the biggest of which seems to be ease of construction compared to traditional web design.I first noticed the shift with community sites like angelfire/geocities and then moving towards social networks, where you just add content.
Now the pendulum is swinging back again because the cost/value equation in favor of corporate networks makes less sense (specifically, we didn't realize the consequences of selling ourselves and our data for 'free' services until it was too late.).
Needed a bit of polish, but definitely promising. Never got the critical mass though.
Understatement! Diaspora was a complete mess.
Security was the problem everyone focused on. Good security is built in at a foundation level and a fundamental component to the entire design and implementation.
Generally when you are talking about a secure application, you have a primitive layer which does authentication and data access, and a layer on top of that which provides logic and user interface with all data access going through that first layer using some kind of authentication token. In this way, a small bug in say, the image upload script, won't let you do much because all operations through the primitive secure layer require an authentication token, which limits the scope of what those operations can do (to say, the logged in user).
Shitty security, like what diaspora had, basically does checks at the top layer (is this user logged in? good.. run this query!). The problem with this is that a small bug _anywhere_ gives you full access to _anything_, which is precisely what was happening. Sure you can patch those small bugs as you find them, but there will always be more.
In other words, it wasn't that diaspora had some security bugs or needed some polish, it was that security wasn't an integral part of the software, which can't really be fixed without a complete rewrite.
The less focused on problem was that the thing wasn't built to a specification, they just kinda started coding it. If you want to build something open and interoperable, that's not how you do it!
And then the main problem was that it had no killer feature to attract users. It did what the other two established offerings did, except without the established user base. Being full of security holes and having no api arn't really thinks most users care about, yet it still failed to gain any kind of adoption.
I honestly felt kind of sad for the team (one of whom apparently killed himself, possibly over stress of the whole thing). They were all very inexperienced, and we've all at some point said "hah, I could write a better <something> in a few weeks!" at that point in our careers. Usually we take a crack at it, realize we are in way over our heads, and it dies quietly. These guys got a shit tonne of attention, were obligated to produce something they didn't have the skills to produce, and then basically crashed and burned before us all.
actually I have to disagree. It is so much easier to stay connected to friends that have long moved away and to reconnect with friends that move back. Of course I am older than most people on Slashdot so yes I have friends that moved away 15 years ago and then move back to town that I want to reconnect with.
It is makes me feel more connected with my brother that lives 3 hours away to see his posts daily on facebook. I do call and talk with him a couple of times a month but with facebook it is daily.
Of course I used to do the same thing with email but email is less popular than it once was.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.