Notes from J. G. Breslin, A. Passant, and S. Decker The Social Semantic Web 3.2.8 Folksonomies Contributed content on social websites may be tagged with.

Notes from J. G. Breslin, A. Passant, and S. Decker

The Social Semantic Web3.2.8 Folksonomies Contributed content on social websites may be tagged with a

keyword by the content creator or sometimes by community members

A tag is a keyword annotation that acts like a subject or category for the associated content

Folksonomies are constructed from these tags:

3.3 Object-centered sociality Social networks are designed to help us work together on common

activities

But many online social networking services (SNSs) lack such common objectives

Social networking sites are becoming meaningless.

And SNSs don’t usually work together

It’s theorized that the longevity of social websites is proportional to the ‘object-centered sociality’ occurring in these networks

I.e. the degree to which people are connecting via items of interest

Some advocate augmented social networks

Citizens self-organize into communities around shared interests

Adding annotations to items in social networks is useful for finding interesting items and people with similar interests

Topic tags, geographical pinpointing, etc.

People’s SNS methods will probably continue to move closer towards simulating their real-life social interaction

Meet others via something in common

Eventually more realistic interaction methods with friends Online connections become intertwined with their real-world

interests

Multiplayer online gaming has had groups (‘clans’) of people working towards common purposes for over a decade

Virtual worlds (e.g., Second Life) have begun to provide a user experience more faithful to reality Networks of friends interact in more realistic ways

10 Social networks10.1 Overview of social networks Anybody is connected to everybody else (on ave.) by 6 degrees of

separation. (sociologist Stanley Milgram)

Cf. the Erdős number, the Kevin Bacon game

Often, one route is followed to get in contact with a particular person

But, after talking to them, there’s another obvious (but previously unknown) connection

Cf. the small-world network theory (Watts and Strogatz)

Even in a small SNS, there can be a lot of links

This data is usually meaningless when viewed as a whole

Usually apply social network analysis (SNA) techniques Can reduce the amount of relevant data by clustering

People modelled as nodes or Relationships (co-authorship, friendship, etc.) as edges

10.2 Online social networking services The idea behind online SNSs is to make people’s real-world relationships

explicitly defined online

Enable communities of like-minded people to group together

Talk about items and interests that they can’t talk about as effectively or regularly via meetups, phone conversations, e-mail

Also useful for networks of geographically-disconnected friends

Valuable for getting in touch with someone whose advice or skills you need, through a friend-of-a-friend or …

Social networks are sometimes used for viral marketing

A key feature of these sites is community-contributed content that can be commented upon by others.

Many SNS are also opening up to web crawlers

Make users’ profiles and content searchable and accessible without being logged onto the site

10.3 Some psychology behind SNS usage Few users realize the very public nature of content on these SNSs

Few grasp the persistent nature of information on the Web

Due to the social norms in SNS-based communities (the perceived closed nature of the community),

individuals may act in a more ego-centric manner than they would in other situations (‘everyone is doing it’)

Virtual reputation is also an obsession with SNS users

A ritualistic element to SNSs

People enter their daily thoughts on a blog linked to their profile and check in for comments regularly

10.4 Niche social networks It’s the fine-grained and targeted communities (e.g., CafeMom,

BOOMj and PEERtrainer) that are experiencing growth

Cf. object-centered sociality above

Suppose the rise in universal SNSs (e.g., Facebook, MySpace) is mirrored by an explosion in the growth of niche SNSs

This would accelerate the demand for semantic-type applications

Let people travel seamlessly through various social networking services, finding objects related to their interests

10.5 Addressing some limitations of social networks Most SNSs usually don’t work together

Re-enter your profile and redefine your social connections when you register for each new site

Standard sign-on systems like OpenID and profile representation mechanisms like FOAF let us define our identity and reuse it

Semantic Web vocabularies like FOAF and microformats like hCard and XFN (XHTML Friends Network) can serve as platforms for

linking or reusing the diverse info about a person from heterogeneous social networking sites

performing operations on such reusable and linked data

‘Social network portability’: the ability to reuse one’s own profile across various social networking sites and applications

Beginning to see distributed social networking platforms where social connections can be formed across sites

Recent developments let people collect and manage their identities across various SNSs (PeopleAggregator, 30boxes [calendaring web application], etc.)

Solutions like OpenID let people have a single sign-on to any of the SNSs they belong to

OpenID is an open standard that lets users be authenticated by certain co-operating sites using a 3rd party service

The DiSo (Distributed Social Networking applications) project aims to implement open-source distributed social networks

Building Wordpress plugins that implement or build on microformats like XFN, hCard, XOXO (wp-contactlist, wp-profiles) OpenID (wp-contactlist, wp-openid-server) OAuth (lets a third-party application get limited access to an

HTTP service)

A smaller set of standard, reusable contact formats could make such services more widespread

The Google Social Graph API can enable this Lets applications reuse social graph info extracted from

sources all over the Web and represented using the open formats XFN and FOAF

10.6 Friend-of-a-Friend (FOAF) Semantic Web technologies allow a more expressive description of a

social network

Use heterogeneous nodes and links denoting different types of objects and relationships

Lets us express a model for an object-centered network

Content and other items of interest are described along with people in a decentralized way

Anyone can create their own FOAF file describing themselves and their social network using tools such as

FOAF-a-matic (http://www.ldodds.com/foaf/foaf-a-matic) or

FOAF Builder from QDOS

FOAF data can be easily produced from social software systems that feature some user profiles and friends lists

Now some examples of tasks that can be done using social network data expressed in FOAF that would be difficult otherwise

10.6.1 Consolidation of people objects Merging identifiers of equivalent instances occurring across different

sources

Object consolidation (‘smushing’) can be done for instances that share the same value for inverse functional properties (IFPs)

10.6.2 Aggregating a person’s web contributions Retrieve content a person has contributed to various sources on the

Web, e.g., all documents, images, chat events, etc.

Hard with a normal search engine

People may share their name with other people

May use different account names on different sites

This query is based on a precise URI—won’t retrieve documents created by the same user while using another URI

One option is to define owl:sameAs statements between this URI and others for the same person

Another is to run the query not based on the URI, but rather based on an IFP

10.6.3 Inferring relationships from aggregated data A sample SPARQL query for extracting the social network formed by

explicit foaf:knows relationships

PREFIX foaf: <http://xmlns.com/foaf/0.1/>

SELECT DISTINCT ?s ?o

WHERE {

?s a foaf:Person .

?o a foaf:Person .

?s foaf:knows ?o .}

Also many implicit social connections on the Web

Links between people may be inferred due to links through some common objects—e.g., people appearing in the same pictures, tagging the same documents, or replying to each other’s blog posts

10.7 hCard and XFN Like FOAF, hCard can be used to define properties relating to

people, including ‘bday’ (birth date), ‘email’, ‘nickname’, and ‘photo’

These properties are embedded within XHTML attributes

Incorporates the Geo microformat to identify the coordinates for a location or ‘adr’ (address) described in an hCard

XFN (XHTML Friends Network) is another social network-oriented microformat

Developed in 2003, just before the creation of the microformats community

Lets us define relationships between people—e.g., ‘friend’, ‘neighbor’, ‘parent’, ‘met’

Supported through the WordPress blogging platform When adding a new blogroll link, use a form with checkboxes

to specify additional metadata regarding the relationship between the blog owner and the person linked to

The different types of person-to-person relationships available in XFN allow richer descriptions of social networks than FOAF (has only ‘knows’)

But FOAF can be extended with richer relationship types via the XFN in RDF vocabulary or the Relationship vocabulary

10.8 The Social Graph API and OpenSocial10.8.1 The Social Graph API The social graph and ‘social network portability’ are mainly about

bringing your social network connections from one site to another

Being on Facebook, you could move to LinkedIn and bring your profile and connections with you

The global social graph is composed of all the social network connections that are distributed across a multitude of sites

Google’s Social Graph API (discontinued in 2012) is a step in this direction

Returns web addresses of public pages and publicly-declared connections between them This data is got from FOAF and XFN info embedded in other

crawled pages Combined with specialist knowledge about the structure of

certain large social websites

Previous FOAF aggregator efforts (Plink, FoaFSpace) didn’t achieve critical mass

10.8.2 OpenSocial The OpenSocial API is an initiative from Google that enables

‘gadget’ portability

Social applications can be deployed across a variety of social networking sites

Google, Yahoo! and MySpace also formed the OpenSocial Foundation (http://opensocial.org/)

A non-profit entity to support such social application portability

The most mature standards-based component model for cloud based social apps

10.9 The Facebook Platform Some of the Facebook Platform technologies available for developers

FBML (Facebook Markup Language) A form of HTML to customize the look-and-feel of one’s

applications

FQL (Facebook Query Language Lets us easily query Facebook data from our applications in an

SQL-like syntax

The Facebook API lets Facebook social data be integrated into applications, e.g. data on friends, groups or events

FBJS (Facebook JavaScript) Lets us ‘sandbox’ JavaScript in their FBML code—the JavaScript

of one application hasn’t access to anything outside of its own scope

All of these have been licensed in more generic forms as

SNML (Social Network Markup Language),

SNQL (Social Network Query Language),

SNAPI (Social Network API), and

SNJS (Social Network JavaScript)

to the Bebo SNS (used to be popular in the UK)

Facebook Connect lets us connect our Facebook identity, friends and privacy to any site

Lets 3rd parties implement and offer features from the Facebook Platform on non-Facebook sites

OWF (Open Web Foundation, established 2008) provides a venue for Google and Facebook to resolve differences between their Open-Social and Facebook platforms

And work on a standard way for their users to interact with each other

The APIs for accessing Facebook data have specific terms of use which restrict the caching of their data

OpenLink developed an RDFizer middleware layer that talks to FQL

Converts the requested Facebook profiles or associated data (e.g. photo albums) returned in XML form into semantic linked data on the fly, without caching

Facebook FOAF Generator returns FOAF data for a user who has authenticated with the service

10.10 Some social networking initiatives from the W3C W3C Social Web Incubator Group (April 2009-Dec. 2010)

http://www.w3.org/2005/Incubator/socialweb/

See the final report

“A Standards-based, Open and Privacy-aware Social Web”

http://www.w3.org/2005/Incubator/socialweb/XGR-socialweb-20101206/

The Social Web should allow people to create networks of relationships across the entire Web,

while giving them the ability to control their own privacy and data

The standards enabling this should be open

The report presents a framework for understanding the Social Web and the relevant standards in the report

Proposes a strategy for making the Social Web a "first-class citizen" of the Web

10.11 A social networking stack Rather than building a separate social networking layer into tools,

info space and application architects need to fold it into various technology stacks

The Nepomuk project (Chap. 13) does this for the desktop

But there’s evolution toward ubiquitous computing and the ‘Internet of things’

Will deliver much more info

The Internet infrastructure itself might need to be augmented to include a social networking infrastructure Keep users from drowning in an ocean of unconnected and

meaningless info

A social networking stack needs to take into account a person’s relevant objects of interest

Provide some limited data portability

The actions and interactions of a person with other users and objects (exhibiting relevant properties) in existing systems

can be used to create new user or group connections when a person registers for a new social networking site or application

Instead of a fragmented view of one’s network in each application,

the social networking stack would let a user use all of their person-to-person connections in any application

Making social networking a shared component across various desktop and Web applications

The cross-application social networking stack requires several layers

1. Personal authentication and authorization layer

Using a single sign-on mechanism (e.g. OpenID, Sxip)

2. Social network access layer

Using the social networking contacts created by an individual across various platforms E.g., by collecting FOAF ‘knows’ relationships from multiple sites

Access control is required: social connections aren’t always bi-directional

3. Content object access layer

Collect a person’s relevant content objects

Verify that they’re allowed to reuse data and metadata from these objects in the current application

For reputation purposes, verify that these items are in fact created by the authenticated individual on whatever sites they reference