Thoughts on the Social Graph

Jamie found an excellent story about the trouble with social graphs. The author discusses the proliferation of social networking websites, the annoying problems this creates, and proposes an open solution to much of the problem. Essentially he is talking about an API for all those relationship systems not under the control of any single commercial entity, coupled with a shared login system. Had things like this been popularized a half a decade ago, we'd be looking at a different internet.

Comments

Yawn.

[morgan_greywolf]

Yet another article about how all social networks should be standardized and have centralized user management. This is the Internet, folks. Decentralization is the name of the game. Get used to it.

Re:Yawn.

[mdwh2]

TFA says: A centralized "owner" of the social graph is bad for the Internet.

It seems to be talking about a system where anyone can run their own server according to the open standard APIs, and hence will not be centralised.

Although he's right that people are tired of readding friends on each network, one flaw is that "friend" has different meanings. On some, it's simply "This person is my friend". On some like Facebook, it also means they can see information about you that others might not. On LiveJournal however (which was created by the author of this article), it goes far beyond simply "friend"; it indicates which journals you want to read, and who can see your "friends only" entries. So conceivably, who I want as a friend on Facebook isn't necessarily the same as who I want as a "friend" on LJ.

Now theoretically this can be handled in that "people whose journals I want to read" could be a subset of anyone I list as my friend (i.e., you have an option for each friend whether you read their entries, whether they can read yours, or whatever is specific for that site). But that's more hassle for individual users.

Re:Yawn.

[Red_Foreman]

I agree, a centralized API would make it far too easy for stalkers. But the article makes a flawed assumption - that people on Facebook want to be connected on MySpace and vice versa, and you (IMHO) can't make that assumption.

Re:Yawn.

[BoberFett]

You can have cross-compatibility without decentralization. In fact decentralization is easier when you don't have vendor lock-in, which is essentially what we have now. A bunch of disparate sites all with different abilities, but no system to easily move from one to another and link any of the data from one to the other.

Re:Yawn.

[makomk]

Actually, it's more about having multiple centralised social networking providers doing different things, and how to keep friends in sync between them. (So you can use Livejournal for blogging, Facebook for keeping track of people you know from college, Twitter for micro-updates, some other site for photo sharing, etc, and it'll help keep your friends in sync between them.) It'll still encourage the growth of centralised social networking sites and still require people to get accounts on several different sites, it just makes it easier for them to do so. As commenters on Brad's journal have said, it's the equivalent of a multi-protocol IM client rather than Jabber.

Re:Yawn.

[mdwh2]

I think that's called HTTP. :-)

That would be a web server though. There are different types of server, e.g., email server, which need different protocols. Sometimes people come up with new protocols (e.g., OpenID).

Re:Yawn.

[Tom]

All the "social networking" sites I know are websites... :-)

Re:Yawn.

[xappax]

I don't want 50,000 social networking sites to know my address...

Social networking is about voluntary information sharing. If you don't want the whole world to know your address and phone number, don't put those things on your MySpace page. If you do, it doesn't really matter which other social networking sites have that information, because it's already public on the internet.

This isn't about scraping and publicizing information that you want to keep private, it's about giving you the freedom to synchronize the information you do want to share between multiple sites.

You already have the ability on most social networking sites to control privacy - making some information public and other bits private, and there's no reason that meta-network tools wouldn't allow you the same control. In fact, they could give you finer control. Perhaps you could restrict a particular journal post to only your friends on Flickr and Facebook, or only show your contact information to people who are your friends on 3 or more networks...

The point is that these are tools to give you more control over your information, not less, so try to not freak out about privacy.

Re:Yawn.

[bar-agent]

Even if everyone supports one standard, the bottom line here is that the data has to centralized somewhere in order to maintain data consistency.

Not necessarily, the data could be distributed, redundant, and synchronized.

IP

[PresidentEnder]

So if I write a little app that I point at my friends page on Facebook or you point at yours on Myspace, which then steals our friends lists and adds them to this wide open free social graph, do Myspace and Facebook have a right to be mad at me?

Re:IP

[mdwh2]

Whilst IP laws can cover "collections of data", it's hard to see it enforceable in this sense. I mean, by the same logic, would they be mad if I recreated the same friends list manually on a different site? Of course not - it's not unreasonable that my friends are still going to be my friends on a different site. I don't see using an automated tool changes that.

Is it an IP infringement if I list my phone number, email and address on one site, then put it on another site too? Of course not. No matter what their TOS might say, I think it'd be hard to argue they own the IP of your personal information, and you don't.

So you want to change the Internet...

[Tucan]

How about an RFC instead of a web page?

To which I respond...

[Billosaur]

...who cares?

Use a social netowrking site, don't use one. Use MySpace, Facebook, or don't. Is this really a problem? No. Is it bothering anyone else? No. Is this news? No. Nothing to see here -- move along.

anti-human

[m0llusk]

People change their social networks all the time. With this in place, you wouldn't be able to just "wash that man out of your hair", but you would have to go online, identify yourself, proceed with authentication, and then click around to make the changes. In addition to big changes in social networks being laborious to enter, their implications grow as well. What about the folks who relied on your network to reach others? Will they give you negative feedback for moving on? This idea seems to be based on a hackneyed understanding of how human relationships grow, evolve, and sometimes just fall away in large numbers. Very young or socially challenged people seem like the only potential customers.

To really try to solve this problem the representation of relationships would have to be automatically generated, and that gets creepy really fast since it would mean having computer applications track all significant interactions with others.

And we want this *why*?

[pla]

What I mean by "social graph" is a the global mapping of everybody and how they're related

Just that? Why, sure, I'll gladly make enough info public on myself and my friends to make identity theft nearly trivial. And hey, as a perk, if I ever find myself on the run from the police (for example, after someone steals my identity and gets me flagged as a major contributor to Al Qaida), they'll have a convenient list of everyone I might contact. Golly, what not to love about that?



People are getting sick of registering and re-declaring their friends on every site

Why, exactly, does "every site" need to know my friends? For that matter, why should any sites know my friends? And I don't mean in the Slashdot Friends/Foe sense - I have plenty of both, solely for the purpose of moderation. Of over 100 people on my lists here, I only actually know three of them, and one of those I've never even met.

If a site actually needs to know my friends/family/coworkers, you can safely bet on my not wanting to use that site.

For the record, I get sick of registering at websites not because it takes too long to come up with fake info, but because for the majority of them, I shouldn't need to create a personalized account in the first place! If I find something through Google, I don't want a lasting relationship with a site, I just want my damned content. If I buy something as a one-off purchase, I don't want an account, I just want the transaction completed and all my info expunged from the site. Unless I specifically ask a website to give me a persistant profile, don't force one on me - it only wastes time, and I won't rememeber what fake info I put in next time anyway (hell, I must have over fifty logins at the NYT).



This sounds like yet another one of those non-issues that give marketing gurus wet dreams and serve no purpose beyond stripping us of any semblance of privacy and anonymity. Brad can keep his thoughts, I want no part of it.

Re:And we want this *why*?

[FiveLights]

Does this mean you won't be my myspace friend?

Re:And we want this *why*?

[mdwh2]

Why, exactly, does "every site" need to know my friends? For that matter, why should any sites know my friends? And I don't mean in the Slashdot Friends/Foe sense - I have plenty of both, solely for the purpose of moderation.

Actually, this is the sense that is meant. Usually sites, just like Slashdot, use "Friend" to imply some specific feature, whether it's who you want to see certain personal data, or whose journals or comments you want to read. It's unfortunate that the word "friend" has been overloaded, but Slashdot is just as guilty of this.

Given that LiveJournal - created by the author of this essay - uses "friend" to mean something specific, I imagine the author is talking about "friend" in this sense, whether or not they really are your friends. Though the flaw I think (as I said in another comment) is that since different sites use "friend" differently, who you want as friends on one site may not be the same as who you want as a friend on another.

For the record, I get sick of registering at websites not because it takes too long to come up with fake info, but because for the majority of them, I shouldn't need to create a personalized account in the first place!

Agreed, and OpenID is one way round this (which was created by Brad, as it happens). This just goes one step further, for those sites (like LiveJournal, Facebook - or Slashdot) where you do want to enter this extra information.

Re:And we want this *why*?

[Colin Smith]

For that matter, why should any sites know my friends? So... The whole concept of social networking has bypassed you entirely?

You truly are Slashdot material. Welcome my brother.

Re:And we want this *why*?

[kebes]

Just that? Why, sure, I'll gladly make enough info public on myself and my friends to make identity theft nearly trivial.

Quite right. And it goes even deeper than that.

I know how to use the web in such a web that I'm "sufficiently anonymous." I know true anonymity is impossible (e.g. with an IP address and a subpoena), but I know how to restrict the information I give out to a level I am comfortable with, and totally out of my control.

One problem with ubiquitously-connected social networking is that I not only have to be careful what I reveal, but I am now very much dependent on what my "friends" decide to reveal about me. If they go mentioning personal information about me, and it's cross-connected through every social networking site I visit, then this represents a release of information beyond what I'm comfortable with.

Obviously this problem already exists (and currently results in, e.g., people wasting time un-tagging themselves from Facebook photos)... but a widely connected and widely available social graph exacerbates the problem. Suddenly I'm dependent upon the net savvy of every single person who is connected to me? (And, given the whole "six degrees" issue, that's a lot of not-so-savvy people.) No thanks.

The end result of more detailed, more available, social information is merely that those of us aware of the privacy implications will stop using social networking sites. Is that really the intent here?

Re:And we want this *why*?

[pla]

So... The whole concept of social networking has bypassed you entirely?

If you mean that I can't call myself one of the 1.4 million "friends" of the latest boy band - Yes, it has. I simply do not see the point of Myspace or Facebook other than as a free-as-in-beer webhost (with the hidden expense of having all your "friends" receive slightly better-targetted advertising).

If, however, you mean a real social network - I limit mine to people I actually know, people that (with very few exceptions) I have physically met. Friends and acquantances whose real names and at least partial contact info I know, whose birthday I might celebrate with them, whose voice I would recognize on the phone or whose face I would recognize in a crowd.



Call me a Luddite, but it disturbs me greatly to think that we have diluted the term "friend" to nothing more than a form of moderation roughly translating as something between fandom and "I like something about your web page".

Re:And we want this *why*?

[ryanvm]

Call me a Luddite, but it disturbs me greatly to think that we have diluted the term "friend" to nothing more than a form of moderation roughly translating as something between fandom and "I like something about your web page".

Hmmm - what scrumptious irony it is that I have added you to my Slashdot friends list because I completely agree with your post.

It already exists - FOAF

[Colin Smith]

http://en.wikipedia.org/wiki/FOAF_(software)

Vendor lockin is the reason it isn't simple to migrate across all the sites.

Re:Security Issues

[kebes]

TFA only briefly mentions the possible downsides of aggregating all of the social networking sites. It says:

It's recognized that users don't always want to auto-sync their social networks. People use different sites in different ways, and a "friend" on one site has a very different meaning of a "friend" on another.

My reaction is much harsher than this. I don't merely want the option of syncing/not-syncing... I would want (ideally) complete control over how widely distributed my "friend-connections" become. Frankly I hate having to maintain all kinds of separate username/password/accounts on different sites. But, I would hate even more if all those different accounts were automatically identified with each other.

The people I communicate with on Facebook are not the people I interact with on Linux forums or on Slashdot. The meaning of a "friend" (or whatever) on each site is totally different. Not only do I not want these connections treated identically... I don't want those separate accounts to be related to one another!

Frankly the downsides to having my online social activity interconnected are numerous: spamming, ease of monitoring me, etc. The end result is that I will either reveal personal information I didn't intend to, or conversely I will use the sites less freely because I'll be worried about revealing information (e.g. if I know potential employers will easily find the information).

Considering the numerous downsides, I have trouble seeing the benefit, to the end-user, of having a comprehensive, widely-accessible 'social graph.'

Not going to happen

[Angst Badger]

"Dear commercial websites, could you please implement a system that will render yourself and your profit models irrelevant?"

It's my understanding that a crack team of programmers has been assigned to this problem. That team includes Santa Claus, the Easter Bunny, and the Great Pumpkin. Good luck and godspeed.

Re:Security/MS Passport

[mdwh2]

Particularly, the use of OpenID. I believe this was the intent of Microsoft Passport (centralized login to all other websites), but I would hardly trust my ONE internet password to Microsoft.

OpenID is decentralised. Being open or a standard doesn't imply centralised (think email - you can email people on other servers, without needing some centralised trustable email server).

Relax, it's just Trillian for Social Networks

[prestidigital]

If you happen to use multiple services then you will see the utility in bringing them all together in one place. If you don't happen to use these services then don't worry about it.

Mugshut

[bjourne]

Mugshot seem to be what he's looking for. It is an open, free software, community, meta site. It tries to create a interconnect all different community sites and place them under one roof so to say. With one centralized user management system. Seems like a very, very ambitious project because it is damn hard to anticipate human behaviour and social patterns. In the broad sense, an internet community is everything from mailing lists to MySpace to Slashdot to various forums and even BitTorrent trackers.

Re:All you need to know on the subject is:

[aftk2]

Ummm... Linkedin? It's been pretty important for a number of people I know..

Comment removed

[account_deleted]

Comment removed based on user account deletion

He identifies the solution is out there...

[alexhmit01]

Facebook has made it possible for people to build the applications that people want and tie them into Facebook. A Web 2.0 site could accept log-ins, or allow Facebook users to simply add the application, adding them as users. Conceivably, Myspace will add a similar feature before going bust. The article's author gives a lot of non-sense about developers not wanting to be slaves to facebook, but they have it backwards.

I subscribe to Netflix. I added a Netflix app to Facebook, it let's my friend's see my queue... yawn... It also let's my Facebook friends, if they get Netflix, quickly add me as a Netflix friend (subject to my approval). The Netflix app mirrors some of Netflix's UI, but not everything. I still go to Netflix to manage Queue's and add movies, but I can see what's going on quickly on Facebook.

The problem is that most Web Developers suck. If your data store for your web-app is good, then you can EASILY create a Facebook front end. If your front-end has all your database calls (no stored procedures in the database, not even a DB functions file in Perl/PHP/whatever you coded in), then you see it as "be a Facebook App OR a website."

The promise of HAVi in the AV world was that we would connect our equipment via Firewire, and they would export a front-end in Java that our TV or Receiver would render for us. The data in MPEG-2 with fixed compression caused content producers to go ape-shit, but the idea is valid on the web.

If you want to process information, you need to collect it and do something with it. The days of a "single HTML interface" are now over. You need a mobile version, an iPhone version (possibly, we'll see adoption rates), and now a Facebook version.

I collect my photos in iPhoto on my Mac. I upload them to Facebook via an iPhoto plug-in to show my friends. I upload them to Shutterfly via an Export Plugin (well, did until they haven't supported iPhoto '08 yet), so my extended relatives can buy pictures.

I have other friends that are into photography, they use Flickr. However, there is a Flickr "interface" for Facebook, so their Flickr Albums are viewable on Facebook. Sure, if they have pictures that they want the Facebook features (tag a friend), they need to upload to Facebook, but if they want Flickr sharing (tags, etc.), they upload to Flickr and put it on the Flickr App on Facebook.

Open APIs will let US aggregate OUR data, not have one site steal it from others.

Conclusion?

[Tribbin]

I always skip to the conclusion before I optionally read the whole article.

quote:

Conclusion:

I'm excited about this. Start thinking about how you can take advantage of stuff like this. It's going to be cool.

How 's that for a conclusion?

Re:what i don't understand about slashdot

[Baavgai]

Good point. However, the pro single sign on, anti universal tracking is not entirely incongruous. It's really a question of control and information revealed.

The vilified universal ID is assumed to be attached to all personal information and controlled by an entity with no particular vested interest in that person's well being. Big Brother bells sound and people start thinking of how to get off the grid.

A single sign on is a little different. It doesn't implicitly involved any information that the individual is unwilling to give up. Sure, anyone could find out all about the activity of a single Internet authority, but, depending on the system, the real person behind the identity is still functionally anonymous. If that ID becomes the focus of harassment, the shielded individual can drop it and move on.

As an individual, I don't like having all my personal information unified for the powers that be to scrutinize. As a tech geek, I'd appreciate using one username and password for all my surfing needs. I don't see this as quite the same thing.