Tens of millions of people have been busy the last few years building Facebook’s most valuable asset - their social graph. As people add friends, and those people add friends, Facebook gets to understand exactly how its users know each other. And as we saw with their “social ads platform,” where users essentially (and sometimes unwittingly) pimp services to each other, it’s not hard to make a little money from data like this.
Google, as usual, is not far behind. But they are taking a much different and more open approach to the social graph. Today they are launching the Social Graph API, which will allow third parties to grab social graph data that is produced by every day activities across the web - linking.
Who you are (defined by Flickr, blogs, Twitter and other web services) and who you know, can be determined by data included with links, or in other data included on web pages but not shown in a browser. The two standards around this, XFN and FOAF, provide explicit and public data to Google (and anyone else that looks) on who you are and who you know.
Technically this is pretty simple stuff. Links may contain XFN tags to state a a relationship, such as “me” or “friend.” These are explicit, public statements of relationships and are built in to many web applications, or can simply be added by humans.
Google is taking the resulting data and making it available to third parties, who can build this into their applications (including their Google Open Social applications).
Third parties are already jumping on board. Plaxo is adding the data to their Pulse profile pages to show additional relationships among users.
Companies can use this data as they please. A simple example is to remind a user of their Google-determined friends, and ask them if they want to add them on the new application, too.
See all
Post
I love how even in Google’s diagram there, they reference Twitter. Shouldn’t they be referencing their Jaiku service/purchase?
This is a very interesting idea - I will need to dig into it more but it may be a big step forward for Data Portability of public relationship data.
@Steve Poland
Twitter is not Jaiku or vice versa.
They are very different services with very different goals
this seems crazy. not sure google will not be pegged as evil on this one. so now one can link all of my comments on various wordpress blogs in one place. this could be powerful in the right/wrong hands.
Jaiku has xfn markup as well. great move my google. This is a very good move forward. Can’t wait to see what this produces
Has anyone tested their Example Applications (http://code.google.com/apis/socialgraph/docs/examples.html)? When I click on any of them I get an xml page saying “Repository access denied.” Does anyone know how I can get them to work?
so what’s to keep me from saying I’m “me” and pointing at techcrunch and exploiting who mike arrington is friends with to gain access to them?
Stay away from Google APIs. Remember when they offered a commerical search API? That was until some innovative mashups started taking traffic away from them.
They’ll take away this API as soon as it matures - just like all the others.
Trev: Try now, they seem to be working.
@graywolf: You sly old “dog”. Your (gray) roots are showing. Always angling…
Talking about Google Social Graph is once again a way to kill the competition’s buzz:
Yahoo Bid: The Buzz War
There i more information about Google and what they can do with the social graph here: http://fishtrain.com/2007/08/2.....ial-graph/
That was a very weak video for something that cool.
i have just changed my name to bill gates now and I live in seattle. please gimme all my friends.
Some of the commenters make a good point above. Whereas Google Pagerank relies on the reputation of other people to assess the reliability of a link, things like XFN take it on trust, so it’s a doddle to spoof. Unless they’re taking account of this somehow; it’d be good if someone from Google could comment.
I know Brad was working with others on this standard months ago. It’s nice to see it finally out, but a little disappointing to see it as a centralised Google API rather than a standard anyone can hook into and host.
As already mentioned above, I’d be wary of using this API, both for long term maintenance of the API and for privacy concerns with your users. The cost:profit ratio to maintain the gathering and caching of this data by Google seems too much in the long run, unlike with web page links (their core business). The easier and more sustainable method to find, for example, which of your friends are already on a site is to upload contacts from your email accounts. And those scripts are floating around everywhere already. “Import your Gmail, Yahoo!, Outlook contacts” as already available on Facebook, Myspace, Plaxo and every other site. Plus, as a developer, you won’t be relying on Google.
All new things have flaws. We will find a solution to the above mentioned problems as they happen.
As a web developer and founder of a social site I’m really interested in these advancements. I personally believe this will be the next major movement (do i have to say “3.0″?). A semantic web can be built off this as a starting point.
This API will stay up. Remember: Google wants openness from YOU not from them.
Remember they used to have a free search API you could use (but wouldn’t even let you use on your own site) because that would cost Google money. OTOH, when they lost with Orkut, now they’re pushing SN openness so they can get all the SN data from MySpace, Facebook, etc.
Pretty simple: when it involves Google helping the net, they’re against open APIs. When it involves everyone else helping Google, they’re all for open APIs.
So simple… Could be great. But it would be also easy to fool this API by producing a lot of spam/junk/garbage links.
I wouldn’t be surprised if my Orku account said:
“One of your friends is already here. He’s name is V1agra.”
regards
s.
@Xavierv: nice conspiracy theory. Unfortunately, given the fact the release was almost certainly timed to coincide with the Social Graph Foocamp which has been planned over the last month, it’s actually Microsoft who had more control over their timing.
just want to know, where is open social?
The way it authenticates that “me” links are really you is that the “me” links need to be reciprocated. For instance, my twitter page says my blog is me, my blog says my twitter page is me, as well as my blog says my facebook is me, and my facebook says my blog is me. Note that this is an associative thing, so your blog could link to all of your other pages for instance, and the other pages need only link back to the blog, not to all of each other. (my facebook = my blog and my blog = my twitter implies my facebook = my twitter).
There are a number of sites now that are aggregrate your presence on the net including mybloglog.com.
Mybloglog includes a very comprehensive set from Bebo, facebook, to technorati and twitter.
All your friends are belong to us. . .
Very cool. I’ve been waiting for something like this. I’ve cranked out a Ruby client library for it already, and it works wonderfully. I look forward to seeing how people use it.
So what exactly prevent me from creating a random page, put some crap data on it and then link to William Gates public profile with a “ME” link ?
In other words - their system assumes that all public information is trustworthy. Therefore it is trivial to poison/pollute the system with intentionally corrupt or invalid data.
It’s nothing more than a proof of concept at the moment, with a very bleak outlook on maturing to the fully thought-through production system.
Do you not understand affiliation? There is nothing in the rest of your “crap data” that would validate the claim of being “me”. You need to understand a bit more about tech before slither through this conversation and make ignorant claims. Read #22 foor a simplified version.
This “innovation” is ready for Beta although it’s not really open or useful like many other Google ideas. Readers have rightly pointed out that Google will suck your info in the name of “Open” but not give back anything.
AlexP: Nothing prevents that, of course, but most intelligent uses of the data will look for mutual links, rather than one-way links.
XFN is hardly a Google invention, and it’s been around for a while.
@kevin so what if you don’t have a twitter page and myspace page I go create them pretending to be you point them at each other and at the your blog which I really don’t own, aren’t you going to connect them all, in some incestous ring? Also what do you do in the case of multiple twitter profiles pointing to one blog? Reminds me of the old show “to tell the truth” will the real Kevin H please stand up …
Certainly very interesting.. now what left to be done is to educate the mass public what the hack this “social graph” thing is…
#27, #29 — It *was* a clueless rant, indeed. I apologize.
@graywolf:
why would I link my blog with a “me” link to someone’s twitter page that’s not mine, but says it’s mine? assuming a = my blog, b = your fake twitter, c = your fake facebook, and x being my myspace page:
a -> x
b -> a, c
c -> a, b
x -> a
You only gain the links in the forward direction, so you don’t gain a -> b by following any links, only that b and c claim to be x as well. However, what I was meaning of only needing to have a single site was more of this topology:
a -> b, c, x
b -> a
c -> a
x -> a
here, you can derive x -> a -> b; x -> a -> c, etc and get the following derived graph:
a -> b, c, x
b -> a, c, x
c -> a, b, x
x -> a, b, c
And hence, all of them claim to be the same person, so the links should be trusted as being the same person, in as much as if someone mentions that the web site b is their friend, then they should be reasonably certain that a, c, and x are the same person.
For a better understanding of “social” search and an introduction to Lijit, check this out. I think Todd knows more about “social” search than Melissa M.
http://www.lijit.com/blog/2008.....ed-search/
So this is a goog initiative to take control of all individuals private relationships on the web …..
If they are so good and relay want users and developers to be in control; why not open source the project ?….oh that would mean that they would no longer be in control….so its not going to happen
My favorite part is when he says, ” Now I have friends, and I’m happy.”
eek…
@kevin I don’t do “engineer speak” so lot’s of your initial google launches leave me scratching my head like I did in high school accounting
The point I’m trying to make is yes you made it very easy for “me” to connect to “me” all over the place, which is the starting place you designed for. However instead assume I’m starting as someone pretending to be you, from a malicious point of view trying to game your system. How much critical mass of fake you do I have to get to make your system believe I’m actually you. The easier you made your system to connect the easier it will be to break through. Also the less of you there is out there the easier it will be for me to become you, digitally speaking.
Social graphs always remind me of those FBI charts of organized crime families.
Be very careful before adopting any Google APIs.
Google is at it’s heart an evil company.
On the one hand, they are working hard to open up social networks (where they suck). On the other hand, they are working hard to drive search towards a closed model.
For example, Google used to have an open web service API for search that you could leverage with your own front end and use in combination with with other search APIs. Now they have deprecated that in favor of providing a AJAX based search component that requires you to embed their widget and branding into your website.
Google needs to provide open access to search engine data - backlink data, page rank, SERP data to webmasters before they can credibly promote any “open strategy”
But of course, as an evil corporation, they are not going to do that.
so when is google going to drop the ‘no’ from it’s motto?
One more thing is the web history from search. if google links those to your social graph as well its almost scary coz people would know what i search!
This is all great discussion, and expanding on #31, the public in general I think will never really care ’bout this social graph nonsense. Not unless perhaps it has celebrities in it. Marketers will care, and perhaps an evolutionary breed of SEO will surface as Relationship Exploitation Optimization (REO) specialists.
But the graph, shiny and new now, will require maintenance and tuneups like an automobile.
Except… this car will never be fully built.
Every day I’m online, I visit web pages with dead links (or the equivalent, a link to a parked domain that once had real content). Unless I’m missing something, the XFN spec does not allow for any dates that would be helpful, such the timeframe of the relationship or the date it was created/last validated. Sites like MyDeathSpace.com showcase the aging and expiration of (social) content, and XFN/microformat relationships will be no different.
Like news-sites’ articles that only link to relevant but older articles, the link trees that are built over time are time-consuming to update for relevancy between all of the nodes (and these are dated nodes).
Anywho… rock on social graph… see you on the inside.
THERE IS NO WAY THAT I WOULD EVEN CONCIDER THIS .
This is indeed a very neat and clever way. If people start sticking to a standard way of linking to other sites with XFN tags then this will really become very useful.
If Microsoft would have done this we would up in arms….
Google does not control the relationships or the api……Just like they dont control any of the content that they spider that does not belong to them…..and turn this content that they dont own into billons of dollars for Google and very little for the original creators of the content….If goog is so good….why don’t they open source the entire project and create a non profit to run “Open” Social….They will not do this because they want to corner the market on the indexing of relationships that should be owned and controlled by the end users and not Google…Will Google ask me for permission to crawl my relationships ?
Can I delete my relationships from the Google servers…The data is on Google’s servers….and they are a company…so they do own the data…
How is this better than facebook ?
Did Google contact any of the other companies that have social networks to ask them for their input ?
No….
Will there be a standards committee that allows end users and companies other than Google to have some input into “Open” Social..Dont hold your breath….its not going to happen…
Any initiative to deal with a “Social” Graph should not be run by anyone company…Google is a company and we should stop fooling ourselves that they are in business for the public “Good”.
Google is in Business to make money and this means that by their nature they will try to dominate with little to no regard to any notion of an open standard that would even the playing field for their competitors.
You are an excellent PR, Michael…
interesting debate on the xfn tags. Valid points raised by people here.
.. I’ve noticed more international sites being integrated into the natural listings.