Will you be my friend?
Have you clicked "no" yet? You know, you get that email that says so-and-so has asked to be your friend at yet-another-social-networking-site.com, you jump over there, and realize you don't really know this person. Maybe they're a friend of a friend you've never heard of, or a link-status junkie trolling for new friends. Do you let your ego take over and bump your social network up a point or two? Or do you turn your back on this stranger, the digital equivalent of averting your eyes when someone you don't know smiles at a party.
Sure, you're my friend. But so what? What possible value is there in codifying my connections to others? I get a nostalgic kick out of seeing who signed my high school year book, but that's about it. And now aren't we doing the same thing online?
I found a glimmer of hope for social networks recently when I stumbled across this procmail script. It's a whitelist generator based on your Orkut network. That is, the script scrapes your list of friends at that site (violating the TOS of course) and uses what it finds to poke holes in your spam filters. So if you're my friend, your email will always get to me, regardless of what your put in it.
What else should my friend-list do for me? Matt Haughey braindumped a bunch of scenarios in an essay a few months ago. Collaborative driving directions, group media recommendations -- all good stuff. I'd like:
- Addressbook My phone number and address should stay current in the phones, address books, and email apps of my first degree, and theirs in mine. And let me request that data from my second degree.
- Media discovery My first degree's preferences (RSS/iTunes most played/Tivo Thumbs-up) should be used as a collaborative filter for new feeds, links, music, and shows in my Now Playing.
- Mobile apps My first degree may ask my phone where I am, sometimes. Much like iChat allows my first degree to ping me, sometimes. Ever play the proximity game with your mobile phone? "Where's the bar?" "Well, where are you?" But I also I want to decide who rings, who goes to voicemail, who goes to email, and who goes to /dev/null. And let's not forget that some folks are more first degree than others.
You see? It's the applications that are interesting, not the infrastructure. And Orcut, Friendster, and Tribe.net are all just infrastructure. They're big databases full of relationships, but very little else. On top of that, the standards are all in place. It's as if these big repositories of data are being handed free APIs: FOAF, iCal, vCard, et al.
The interesting thing is that the social network sites own the contract between you and your so-called-friends. Sure, I can publish a FOAF document saying I know Bill, Tim, and Oprah. But that's meaningless, of course, unless Tim, Bill, and Oprah also agree. It's services like Orkut that aggregate these agreements.
For example, when I plan a trip, I try to find out who else will be around so I have people to hang out with. So my calendar should ask Upcoming.org, "Hey, Jeff says he's friends with Tim. Will he be in New York for GEL?"
Ongoing should say, "Well, let me check. Hey Orkut, are Tim and Jeff friends?"
And Orkut should reply, "Yeah, Tim says he knows Jeff. Go ahead and give him access to Tim's travel calendar. Oh, and here's the email address Tim actually reads."
Now, hook up the same pipes to AudioScrobbler, del.icio.us, Mailblocks, BlogRolling, and any other site that connect me to others. This, by the way, is more or less what Amazon, Verisign, and your Web browsing software all do when you type in your credit card data. Two distinct parties ask a trusted third-party to validate a relationship.
Sure, it's a lot of work -- I've worked on committees developing standards before. It's very, very difficult to get the kids to play together. But the pieces are all there. And the first one to connect the dots will win...
This entry was written by Jeffrey Veen and posted 29 March 2004 at 10:50 PM. It was filed under Technology.
Yeow. These are all kickass ideas. Hopefully someone pulls a string between all these services someday soon.
You have hit the nail on the head. This is a large chunk of the Personal InfoCloud, which will get written-up in one of our favorite web-based journals in the next month or two.
It is all about using the services and repurposing the information to fit the user's needs. Most sites only focus on getting the information in front of the user and not helping the user reuse that information. Yes, date information should dump into my calendar. Yes, my calendar should ping the conference site to see whom I may know that is also attending.
I completely agree, I don't believe we need another portal to manage our information, but our information should be centralized for our own use. We should be able to categorize the information as we need. We should be able to move the information around in our personal infocloud so that it will synch to our phone/PDA from our centralized repository. Having the information at hand (in hand) when you need it is very important. Benefitting from other's shared information to help us better make decisions is what we are all desiring.
Let's make it happen. We have all been dreaming too long and it is time to get it done.
I have yet to use any of these social network services. Why? Because I see no useful purpose. However, if these ideas were implemented, I might just consider them. (Are you listening?) I'm sure I'm not the only one who feels that way. That sould be incentive enough to implement at least some of these suggestions.
Excellent ideas Jeff.
I think most of the people who blog about social networking miss most of the point.
What do you get from codifying your friendship in a computer with a friend ? Nothing. What you do get the possibility of (based on your friend's permission) to possibly connect with their friends, for reasons that make that real-world friendship important.
For example, let's say that I'm looking for a job. I may or may not know a friend of a friend is looking to hire someone like me, and besides having a mass emailing among my friends and relying on their memory, or bump into the guy at a party, I might not ever know it, and miss out on a job.
In the real world, vouching for a friend gets you pretty far, at least with things like dating, hiring people for a job, and selling or buying certain things. I might not care who I sell my mountain bike to, but I might care about who I rent my vacation house out to.
Social networks can have great value, if they are used right.
Good thinkin', Jeff! You're right - it's about the apps, not the infrastructure.
And yes, I hope you'll be in New York for GEL :)
You've touched on many of of own thoughts when it comes to social software. There are some great ideas in there. For now I'd be happy to just have one service to deal with or something that tied them all together. I've actually gotten some good use out of both LinkedIn and Friendster, but I hate to have to maintain both.
Nicely done anyway. Here's to hoping some of your ideas become reality.
Connecting the dots ... tying it all together kind of reminds me of the same issues that people are facing in an attempt to organize federated identity management.
Wow, both complexity and simplicity.
A person may be only one or two degrees away from you, but requires half a dozen API's cross-referencing each other to confirm it.. awesome stuff Jeff.
We're already seeing what some talented folks can do with mailing list visualisation tools ( http://www.octapod.org/adam/mt/archives/000710.html ), and this awesome work may be but one slice of relationship mapping pie.
Great post! Last month, I posted about what I see as the "killer apps" of the next generation of social software:
http://www.tallent.us/CommentView.aspx?guid=41c1874a-dcfc-4a4a-9eaa-845ec863b6da
The procmail script intrigued me, since I've been thinking along a similar vein about how to use a "6 degrees"-type concept to create an automated whitelist that includes everyone I could ever *want* email from (i.e., not just people I know, but people who travel in the same or adjoining circles) while avoiding people who are almost certainly spamming me:
http://www.tallent.us/CommentView.aspx?guid=6b066143-069b-486d-a12d-13ef6d12189f
http://www.tallent.us/CommentView.aspx?guid=0faa8e94-9ad7-41bb-b7ea-6cf8c6ec65bf
http://www.tallent.us/CommentView.aspx?guid=d9c94422-37f9-4ddb-9928-ac6e4d5fa70b
http://www.tallent.us/CommentView.aspx?guid=087b1aa8-ae0a-4f3b-981a-c7b5bace4ae4
http://www.tallent.us/CommentView.aspx?guid=68d7222e-4e08-4fbf-a064-503d49822f47(hehe... I really need to get a wiki going and condense those down to a single argument).
I think FOAF "knows" is not terribly useful, but search engines that map out XFN data (like rubhub.com) can show much more interesting relationships, or lack thereof.
With all due respect, the title of your post is "Will you be my friend?" (asking yes or no), and yet, how is this any different than "Will you be my foafknows?" (asking yes or no).
The key to some of the applications you mention, like addressbook, is recognizing that perhaps people want to share different information with different networks.
E.g.:
Addressbook, perhaps some of it I would share with a *friend*, and perhaps other parts of it I would share with a *co-worker*, and perhaps a few more parts with *family* relations.
Media discovery, perhaps I want to only collaboratively filter using what my *neighbor* is listening to, in the hopes that maybe I'll discover a cool local band.
Mobile apps is another perfect example of sometimes you want to be in touch with the work crowd, sometimes with *colleagues* at a conference, sometimes with friends, and sometimes only with that special someone.
You mention yourself that an application may need know that "some folks are more first degree than others", like a *spouse* or a *sweetheart*.
This is exactly what XFN [ http://gmpg.org/xfn/ ] let's you quickly and easily do. XFN allows you to make the general distinctions among your relationships which would enable the sorts of applications that you speak of to provide much richer functionality. For example the *emphasized* terms above are all already defined in the XFN Profile [ http://gmpg.org/xfn/1 ].
In fact, you can see for yourself that a lot of this is already in wide use by checking RubHub [ http://rubhub.com/ ] , the XFN aggregator. You can even see that you yourself are already a part of the XHTML Friends Network just by checking your Rubhub profile:
[ http://www.rubhub.com/?xfnFind=http%3A%2F%2Fveen.com%2Fjeff%2F ]
netscreen
Currently:
() More...
About Me
Bio: Jeffrey Veen
Book: "The Art & Science of Web Design"
Book: "HotWired Style: Principles For Building Smart Web Sites"
Work: My LinkedIn Profile
Travel: China, Tuscany, Kayaking in Baja, Touring Costa Rica, Studying Theater in London
Popular Posts
» Making a Better Open Source CMS
» Seven Steps to Better Presentations
» A Contrast in Urban Design
» IA Jargon Watch
» On Writing Short
» Pain and Cycling
Recent Photos
XML Feeds
Subscribe to my site
Click the link above to be notified automatically every time I add a new post.
