Re: What is a social network?

On Tue, Jun 9, 2009 at 9:51 PM, Melvin Carvalho<> wrote:
> On Tue, Jun 9, 2009 at 3:53 PM, Christine Perey<> wrote:
>> I was asking myself, when populating the list of 30 popular community
>> services: What defines (for those of us in this W3C XG) a service as being
>> unequivocally on the list and why?
I more or less was intuitively choosing the sites as social networking
sites, not just Social Web sites. I imagine in the future this sort of
thing will become more and more blurry, as Google Wave demonstrates.

 I chose sites where there was some kind of explicit social network
that featured some sort of login and profile, and allowed you to make
a list of friends, somewhat similar to Boyd's definition.  Many of the
sites were a bit confusing in this regard, given the fact that they
weren't in a language I understand.

It was a first-stab, and as I believe all or less *all* information on
the Web is social, so just listing *all* websites as social networking
 sites would be absurd.

I'm noticing that the overlap between the list from Alexa and
Christine's list is pretty high, at least 50 percent.

>> Do we all have the same definition of social networking? Judging from the
>> services on the first list on this wiki page [1], probably not.

Please add sites that you think should be on the list to the wiki
page! Concrete examples help a lot in this area.

>> I suggest that we take a stab at a W3C Social Web XG 'accepted' definition
>> of scope.

Good idea.

> 1. A system of 2 or more agents (usually associated with a primary
> key) that allows connections to be made between agents.

Should the connections be explicit and Web-based? I think so.
Otherwise, every e-mail is a social network. Yahoo! and others note
that products like Yahoo! Mail are an implicit social network, but
even then they try to roll out explicit (like the failed Yahoo 360
service) social networks.

> 2. The agents tend to have a primary key, the four biggest, each with
> over a billion instances, are: ( phone number, email, URI, Instant
> Messager ID ) though some of these tend to be open to attack by
> spam/spim.  Perhaps URI is the most spam resilient.
> 3. Strictly speaking, mathematically, the connections in a "network"
> as opposed to a "graph" can have a weighting (but no one seems to use
> this)

Lots of sites have friend rankings (as disturbing as I think this
is!). That's a sort of ordinal weight.

> The phone network and email are good examples of systems where you can
> talk to agents in other subsystems, social network websites tend to be
> more insular. but seem to be evolving towards a more open model.
>> Playing the devils' advocate here:
>> One of the problems with a public/published W3C definition of "social web"
>> or "social networking" is that it may need to be examined regularly and
>> modified/updated. Where will the boundaries every end?

I don't think there is a definition. However, we do seem to have an
ability to spot them, as the fact that our lists are very similar

>> Social networking ("community") features are beginning to permeate many
>> other (previously non social) digital services. There are already and will
>> be more social "features" added to everything.

Yes, but then saying "Hey, everything is social" really doesn't help
anyone. What I think our role could be would be help provide guidance
for Web services that want to add these types of social features: Show
them how to add these features  in a open, extensible,
privacy-enhanced, accessible, internationalized and Web-based way.

>> Some examples:
>>  + a music (or any entertainment content) service where people rate the
>> tunes, movies, etc
>>  + a weather service in which you can see the weather near your friends,
>>  + a local nightclub search service on which you see where your friends have
>> already congregated,
>>  + an auction service on which your friends or taste neighbors offer
>> opinions/advice before you conclude a purchase
>> Take this out into the distance and what you have is an infinite list of
>> socially-aware services. Everything is social...
>> [1]

See above comment on why saying "everything is social" may be true,
but not the most productive way forward.

>> Christine
>> Spime Wrangler
>> mobile (Swiss): +41 79 436 68 69
>> from US: +1 (617) 848 8159
>> from anywhere (Skype): Christine_perey

Received on Wednesday, 10 June 2009 05:24:11 UTC