The Problem with "Google's Social Network" isn't "Social Network", or even "Google" -- it's "'s".

[jducoeur]

[I posted this to G+ a little while ago. It occurs to me that people here might also be interested.]

Specifically, it's the possessive.

Google+ isn't really the big problem nowadays. Nor is Facebook, nor Friendster, Orkut, LiveJournal or what-have-you. The problem is walled gardens. The problem is that all of the big social networks have been designed in a way that is fundamentally at odds with the spirit of the Internet, and they are causing problems for precisely that reason.

I mean, really -- we don't talk "Google's Internet". We don't talk about "Facebook's World Wide Web". We don't talk about "Comcast's email". We used to do that -- back in the 80's, we called those things Prodigy and AOL. You will note that those largely bit the dust a long time ago, and that was for good reasons: because an open system, built entirely on open protocols, with highly distributed and negotiated authority, just plain works better for its users in the long run than a system that is controlled and bottlenecked by any one company or government.

The nymwars are a sideshow -- a fine illustration of the problem, but not the problem itself. The real problem is the fact that the nymwars are even possible, that a company is able to control your social network that way, that your only alternative so far is to go look for another walled garden to hand your control to.

It's time for this nonsense to stop. The solution isn't to build another site, or even another network. It's to focus on the protocols of social networks, and how to make them work together -- to take ownership away from any one group or even a small number of them, but to declare that we want to work in a larger system that forces those companies to compete with each other and work together on the same turf. And if the big companies refuse, to build those open systems and render those companies obsolete: leave them in the dustbin of history, alongside all the previous firms that insisted that their way was the only way.

It's time to take back the social network. Our social network.

Pass it on.

[Afterword, specific to LiveJournal: LJ is, it should be noted, somewhat better. The system is designed to allow both friending and commenting with folks outside the LJ network per se, which is a big step over the big players. Still, my conclusion from this whole mess is that the underlying assumptions are corrupt: that we need to demand a proper separation of identity, network and applications, view them as a true protocol stack, and rethink how it all works. Whether LJ is helped or hurt by that will depend a lot on how it plays out. If BradFitz was still running things, I suspect he'd be spearheading this effort and LJ would be leading the charge; as it stands, we'll see.]

Comments

[be-well-lowell.livejournal.com]

"On the Internet, nobody knows you're a dog."

I have some technical thoughts, too, but this seemed a more important comment for the moment.

[laurion]

I suspect we'll not see this happen unless one of the open source projects takes off. Corporate entities have no interest or impetus to build open systems. They benefit by being the biggest walled garden around. There is no advantage to them figuring out the complications in making a system work across disparate implementations.

I can't imagine, in theory, that building a properly distributed open system is significantly more complicated that building a properly balanced multi-server-farm closed system, but any such system has to tackle the network effects. Oddly, Google+ coming on to the scene with such force may actually make it easier for a new solution to penetrate the market.

I'll know where finally there when Comcast operates it's own set of Diaspora (or whatever) servers for it's customers, the same way it runs email for them, etc. (But please note the overwhelming number of people who have traded in their ISPs email for a centralized service like Google's.) Or when the most popular social network is built on an open codebase that anyone can practically, and interoperable run from their own web host. Wordpress.org/.com fits that model for blogs, but that's the closest example I can come up with. And as with email (or Jabber), there's nothing stopping Google from adopting the open structure, integrating it into Google+, and continuing to operate it under their own policies, while still attracting a plurality of the userbase.

[jducoeur]

I suspect we'll not see this happen unless one of the open source projects takes off.

Oh, of course not. I'm really stating principles here -- the strategy and tactics are far from simple. Moreso because I'm fairly sure that you can't get rich from building it correctly, almost by definition, and the big companies are mainly focused on getting rich.

There's a lot of Just Plain Hard Work involved here: getting The Real Social Network up and running on multiple open-source systems, getting some smaller-scale hosts profitable on the back of that, and slowly peeling groups away on the basis that the real network is *better* than the walled gardens. (In various respects, but let's not kid ourselves -- it stands absolutely no chance of happening unless it can demonstrate that it is, in ways meaningful to at least some people, a better option.)

Only once that happens, and is gaining some real traction, is there even the slightest hope of the major networks playing ball. Indeed, until smaller networks like LJ and DW are getting seriously into it, I wouldn't hold out any hope at all of getting Google's attention, much less Facebook's. I'm assuming it's a multi-year project at best. (My guess is five years to get to the point of leveling the playing field -- a very long time in Internet terms.)

I can't imagine, in theory, that building a properly distributed open system is significantly more complicated that building a properly balanced multi-server-farm closed system

Actually, it's more complicated from a network-stack perspective. Doing this right isn't just a matter of building a farm for a single product -- it's designing at least three separate stack layers, and I believe probably four: identity (partly solved problem); network (several stabs at a solution, no clear winner yet); aggregation (pulling together multiple identities into a single sign on in a secure way); and application (protocols for exposing just enough network information to allow apps to work on the network without demanding privacy-breaking amounts of data).

Part of the point here is that, while a single system *can* implement all of these (and many probably will), they can and should be built so that they don't *have* to be overly unified. I should be able to build CommYou on top of the overall social network, without running a social network myself. Someone who needs hard privacy guarantees should be able to run via a secure provider, but still be friendable (as a pseudonym) by someone on a completely different set of providers, and communicate via agreed apps.

(But please note the overwhelming number of people who have traded in their ISPs email for a centralized service like Google's.)

Oh, absolutely. And mind, I don't really expect that to change. Indeed, I don't even want it to. The name of the game here isn't to put Google and Facebook out of business -- it's to break down the walls and claw ownership of our networks back under our control, so that I can take or leave Google without needing to completely duplicate my flist.

Basically, this is about turning the social network into a fully-interlinked commodity. The big players will fight that tooth and nail -- they want their walled gardens in the worst way -- but it's probably necessary, and likely possible. The result still potentially leaves Google and FB as the biggest players, and that's fine, so long as the walls come down and you don't have to choose between them in order to talk to your friends...

[fairdice.livejournal.com]

Speaking only as myself here: I think you're wrong about Google's desired end-game. If social networking had developed as an open stack to begin with — that is, if the content and identities and relationships had been the sort of thing Google could crawl and learn, like the rest of the web — then I doubt Google+ would exist.

But without a substantial challenger to Facebook, I don't see any hope for things moving in the direction that either you or Google would prefer.

[umbran.livejournal.com]

I have to agree here. Google traditionally hasn't made money off of making a system closed. They've made money off of being able to show people ads.

If social networking became decentralized, so that what the purveyor sold you was a client, and Google could put targeted ads in that client, they'd probably be happy.

[jducoeur]

I entirely agree. But that's why I find the real names policy so strange (at least, their apparently firm attachment to that policy, despite wild objections both inside the company and out), and why I'm having to look around for motives: it seems deeply counter to Google's apparent interests. When I see a company like Google doing something that appears highly irrational, I'm forced to ask why...

[umbran.livejournal.com]

Real names are in line with the advertising motive. It allows for a much stronger ability to profile and target advertisement. If they want to take the step from offering a social network service, to having an "identity service" to being able to do commerce with that identity, real names makes a goodly amount of sense.

Imagine Amazon's "one-click" shopping, all over the internet, based on Google's identity service for both the shopper and the seller. That's a target that calls for real names.

[jducoeur]

Maybe -- although only if enough people play ball. The falloff in traffic on G+ since the nymwars started has been steep, and I'd bet that the premium for real names per se is relatively modest in practice. So this may well be a poor tradeoff.

The premium for providing an *identity* is probably substantial, mind. But as I've pointed out repeatedly, I'm a fairly good example of why the real name argument is kind of tangential here: the *vast* majority of my online identity is under my nym. (By a factor of 10-1.) So providing my real name is actually far *less* useful from a marketing POV than providing my nym: it actually hooks to much less information. And I'm by no means unusual in that: most nymish people I know use a relatively consistent nym across many services.

Granted, you can fairly easily get from one to the other in either direction in my case. But it is vastly easier to suss me from my nym than my name. Especially for marketing purposes: my nym exposes much better what my *interests* are, and that's where most of the discretionary money goes.

I will reiterate, BTW, that I am not arguing in favor of true anonymity. Anonymity is simply incompatible with a good social network. I'm arguing for pseudonymity -- persistent identities that don't necessarily have to be wallet names.

Also, this is a fair argument for Google *having* your real name, but a very weak one for *displaying* the real name inside G+. Indeed, a typical advertiser likely doesn't care quite as much about even receiving your real name as receiving Google's analysis of, eg, your demographics and interests. (They wouldn't object to it, I'm sure, but a Google-driven analysis is gold.)

Additionally, it's worth noting that even Gartner -- source of much of the hardest-edged and most careful analysis of business in the industry -- seems to think that this policy is a bad idea for Google. Granted, this isn't an official statement, but Gartner bloggers do *not* speak casually as a rule, and are rarely swayed by squishy arguments. Gartner is all about taking a cold and impartial look at a company's business prospects.

So I'm skeptical about this line of argument. It's been advanced frequently, and it's a fair reason to *encourage* real names. But the hard-assed policy Google's been applying is counter-productive to this, since it drives away potential targets with no benefit.

And it is worth noting that I have never heard this argument advanced by anyone inside Google, even anonymously, despite how frequently the idea comes up from outsiders. That doesn't prove anything, but there's been enough leakage from Google otherwise that the absence is notable. (Whereas the idea of "government compliance", in various forms, has come up repeatedly, both publicly and privately...)

[jducoeur]

Could be -- as I've said, I'm guessing here. The thing about the identity-service strategy, though, is that it kind of makes sense, and even motivates the real names policy in a way that isn't *entirely* daft. (I think that the logic is incorrect, but not entirely loopy, which is an improvement over most theories I've heard.)

But at this point, I'm still on the outside of G+, barely able to even look in, and am therefore in the unfortunate position of having to guess at motives. As I've mentioned, it wouldn't be terribly difficult for me to be swayed here. But I honestly believe that, unless Google embraces an open social-network architecture, publicly and fairly aggressively, it is likely to lead the company down some fairly evil roads, simply because it is pretty hard *not* to do so once you have the walled garden set up.

That is, I agree that having a solid FB competitor is a good thing. But that competitor needs to be *architecturally* better in order to really improve the situation for people on the ground, and from the outside, G+ isn't. It could be -- but based on what's been said so far, it looks as bad or worse.

Seriously: if all Google is trying to do is have a social web that it can crawl and learn from, it *should* open things up. I sincerely believe that an open-network initiative *led* by Google, instead of fought by it, would knock Facebook to its knees relatively quickly -- that's the kind of genuine game-changer that remakes the world fast. But the real names policy is *completely* in opposition to such a model, which is the primary reason why I find it so concerning. If Google really gives a damn about that policy, then it can never open up...

[mneme]

Very true.

Apropos of which, are you on Diaspora yet?

[jducoeur]

Not yet -- I need to get myself signed up. Not necessarily right this minute, though. One of the FAQs I've been getting is, "why don't you just join Diaspora?", and my response is a little nuanced because I care about the overall architecture of the ecology -- so it's all about protocols and interaction, not code.

Diaspora may well be doing everything right, and kudos to them if so, but I need to sit down and lay out my target architecture before I get too deeply involved there, so that I can make sure I see the forest for the trees. The ideal situation would be that Diaspora already implements the right architecture, but I need to know what that means first...