W3C Workshop on the Future of Social Networking
Agenda 15-16 January 2009, Barcelona

The discussions of the workshop were fed by the input of the 72 position papers submitted by the participants, and animated by the Program Committee composed of experts from the industry and academics on this topic. See below for the 3 main topics of discussions identified by the Program Committee .

Thu, January 15

Time Topic
8:30-9:00 Registration, logistics
9:00-9:20 Welcome and agenda review, by Christine Perey [slides]
9:20-9:30 W3C Overview, by Dominique Hazael-Massieux [slides]
9:30-10:20 Presentations on Appropriate Architectures for Social Networking:
  • The Social Web: Small Businesses / Big Solutions, by Timothée Anglade, AF83 [slides, paper]
  • Managing Social Communications Identities, by José M. Gonzalez, Telefonica [slides, paper]
  • Leveraging social data with semantics, by Fabien Gandon (INRIA) [slides, paper]
  • Current Issues with Social Networks representations, Peter Mika, Yahoo! Research [slides, paper]
10:20-10:45 Break
10:45:12:15 Breakout session on Distributed Social Networking and Data mining and the use of semantic web in social networks
Questions under consideration:

What is the possible role of W3C, and what should the W3C be sure to avoid?

Distributed Social Networking , animated by Dan Appelquist (Vodafone)
  1. Can a decentralized architecture be sustainable, profitable, and usable?
  2. What do we stand to lose if we adopt a decentralized architecture?
  3. How would you prevent fragmentation of web capabilities, and how would that affect standards definition?
  4. What is the minimum set of new functionalities that the future web should incorporate?
  5. How can we allow users who may want to deliberately fragment their online identity to do so?
Data mining, animated by Peter Mika (Yahoo!)
  1. Is a common data format needed for use-cases like social data integration and data-mining ? Or can we just use a common API?
  2. If so, what are the characteristics of the common all-purpose data format?
  3. Of current options (a common API, XML-based data format like Atom, JSON, RDF, microformats), what are their advantages and disadvantages?
  4. To what extent should domain-specific data, such as tags and addressbooks, be standardized, or can we extend in a decentralized manner?
  5. How can other parts of the needed technology, such as privacy, be handled by this common data-format?
12:15-13:15 Lunch
13:15-14:00 Presentations on Privacy and Trust and Distributed Architectures and Business Models:
  • Enabling Trust and Privacy on the Social Web, by Alexandre Passant (National University of Ireland, Galway) [slides, paper]
  • Towards an OpenID-based solution to the Social Network Interoperability problem, by Davide Palmisano (Asemantics) [slides, paper]
  • A Telecom Italia view on the future of Social networking, by Claudio Venezia (Telecom Italia) [slides, paper]
14:00-14:15 Movements to breakout
14:15-15:30 Breakout on Privacy and Trust and Distributed Architectures and Business Models
Questions under consideration:

What is the possible role of W3C, and what should the W3C be sure to avoid?

Privacy and Trust, animated by Axel Ferazzini (OMA)
  1. How do we educate users and publishers on how to manage identity fragmentation, privacy and trust issues?
  2. What are the concrete problems around privacy (e.g., confidentiality of information) and how are the similar, the same or different for those problems associated with trust (determining trust of data, reputation) in social networks?
  3. How can the above problems be addressed in social networks with existing technologies, how can these existing technologies perhaps be combined, and should they be extended or are new protocols necessary?
  4. What problems are not yet coped with by current privacy and trust technologies?
  5. What are the legal repercussions of these sorts of technologies and are new legal or policy frameworks necessary to deal with social network privacy and trust?
Distributed Architectures and Business Models, animated by Marcus Ladwig (Peperoni)
  1. What are the business models for community services in scenarios where communities are less centrally controlled (and/or decentralized)?
  2. What are the roles of mobile network operators in "open" (opposite of walled garden) social networks and how can they produce value by brokering among various ecosystem's partners who generate revenues commensurate with their contributions?
  3. Metrics: How does the industry itself (and others) measure the value and other parameters of social networks [question is for centralized as well as decentralized scenarios]?
  4. How are social networks, in particular on mobile networks, changing/creating new value chains and how could these affect the creation of future products/services and influence their lifecycle?  
  5. What is missing and what is necessary (anything the W3C can do?) to improve and streamline the domain of transactions (billing, trust, payments, micropayments and picopayments) within social networks?
15:30-16:00 Coffee break
16:00-16:45 Presentations on Deeper and Adaptive User Experiences
  • Trends in mobile social networking for mainstream consumers and supporting technologies required, by Lisa McKnight (Nokia) [slides, paper]
  • Social networking across devices: opportunity and risk for the disabled and older community, by Henny Swan (Opera) [slides, paper]
16:45-17:30 Plenary discussion, animated by Dominique Hazaël-Massieux (W3C):
  • Henny Swan, Opera
  • Lisa McKnight (Nokia)

Questions under consideration

  1. How can we more deeply integrate virtual communities (reduce the barriers to access, make communities more useful) into daily life? What is a good format (could we build a "guidelines" for mobile community) for the interface for the consumer?
  2. How is the welcoming door created to encourage people not only to enter once but to return to social networks regularly?
  3. Do users want a high (or variable) level of integration/adaptation? What are the consumers' level of awareness, expectations and needs with respect to service adaptation?
  4. What types of adaptation can be done (1) on the device (automatically) (2) by the network (automatically) and (3) by the user (manually)?
  5. What are the trade-offs to each of these three areas of automation and adaptation?

Fri, January 16

Time Topic
8:30-8:45 Agenda review, summary of previous day discussions
8:45-10:00 Reports from break out sessions from previous day
  1. Report on Architecture for Distributed Social Networks, by Sam Critchley (GyPSii) [no slides]
  2. Report on Semantic Web and Data Mining, by Peter Mika (Yahoo!) [slides]
  3. Report on Privacy and Trust, by Karl Dubost [slides]
  4. Report on Business Models for Decentralized Social Networks, by John Kemp (Nokia) [no slides]
10:00-10:20 Break
10:20-11:30 Presentations on Context and Communities
  • Context in Social Networking , by Julian Pye (Vodafone) [slides, paper]
  • Sentient Computing meets Social Networking, by Simon Hay (University of Cambridge) [slides, paper]
  • Capturing, Using, and Storing Users' Locations, Dan Melinger (Socialight) [slides, paper]

Suggested reading: The Tangled Web We Weave, Greg Howard, Rajesh Kuppuswamy, Kaushik Sethuraman, Microsoft Corporation.

11:30-12:30 Panel discussions, animated by Sam Critchley (GyPSii)
  • Simon Hay, University of Cambridge
  • Julian Pye, Vodafone
  • Dan Melinger, Socialight

Questions under consideration:

  1. Beginning with the user as the focus, what control should the user have (minimum levels of control, optional controls, etc) on how his/her context is used and by whom?
  2. Time is another aspect of context. How are a user's community history (messaging history, travel history, search history) to be used in contextual services?
  3. What formats are contextual data captured and stored?
  4. What would the ideal/best API for context data include?
  5. What are the privacy/security measures which should be in place and what are the responsibilities of any service which is a provider or "consumer" of location/context data?
12:30-13:30 Lunch
13:30-16:00 Next steps
  1. FOAF+SSL: an implementation of a decentralized Social Network, by Henry Story (SUN) [presentation, paper]
  2. Proposed Social Web Incubator Group, by Harry Halpin (Univ. of Edinburgh) [slides]
  3. Conclusions and actions from Workshop, by Dominique Hazaël-Massieux [slides]

Topics of discussions

In reviewing the submissions to the workshop, the Program Committee identified 3 main topics of discussions.

Appropriate Architectures for Social Networking

Problem Statement: When, as is the case for most virtual communities today, a service is operated as a business, for profit, the operator of that community and its business partners make promises to the community members and in order to deliver on these, must carefully manage the resources in and around the community. The operators of communities have also invested considerably in the development of their platforms and prize the uniqueness of their services and the special relationships they have with the participants of the community. While the one-vendor model of social networking has many benefits and will likely persist in the future, many of the position papers submitted for this workshop's consideration are calling for massive decentralization of community (technology, user, business) silos (as currently designed, with centralized control).

The authors of W3C workshop position papers argue that centralization slows down/prevents the development of healthy partner ecosystems, limits the user(s) ease of access to one another (low connectivity between walled gardens) and has other negative effects.

Distributed Social Networking

Digital experiences are increasingly becoming "social". To make it as easy for users to seamlessly control and share their social data, there need to be fewer impediments. There are several initiatives, both industrial and academic, aimed at developing a decentralized infrastructure for social networking.

A number of difficult issues including authentication, storage and ownership of information, trust, privacy and encryption, APIs, integrating diverse data formats, and the dissemination of information, need to be considered. There is also a wide desire for some sort of standardization in this area, with many groups outside the W3C already doing work, such as the Open Web Foundation, DataPortability.org, and the DiSO Project, and many already implement solutions like OpenID in this space. This session consists of several initiatives, both industrial and academic. Furthermore, we will discuss how one or more decentralized infrastructures for social networking might interoperate over the entire Web, including both desktop and mobile environments.

Data mining

A number of different kinds of data, ranging from the output of APIs to various domain-specific formats, needs to be mapped to some common data format in order to enable social network analysis and data-mining. Microformats, XML-based formats, and common APIs must all be able to be mapped to each other to enable data integration. There are many different kinds of social data need to be mapped together, ranging from contact data in vCard to social network data in XFN. Outside of strictly social network data, there are many types of content, ranging from tags and domain-specific content, that should also be capable of being connected. One simple method that has gained increasing traction is standardizing on a simple common API, but a single API seems unable to handle the diverse and ever-expanding types of data available on social networks. While Semantic Web-based technologies could be the solution to this, it are too complicated and heavy-weight. This session will compare the various options and determine what the major problems are with each option, and then try to determine to what extent domain-specific data for social networking should also be standardized.

Privacy and Trust

One aspect of the future of social networking is privacy and trust, and this becomes even more important in decentralized systems. While many people think that trust and privacy are simply relics of a bygone era in social networks, for many social networks some sort of trust and privacy requirement is important. First, in a decentralized environment, a user would need to be able to use a single login-in to access multiple services, and technologies like OpenID already exist to solve this problem. The reverse of identity aggregation must be also be considered: it can be equally important for many users to keep their identity separate over social networks. Second, what other users, developers, and operators, does the user trust for services and to whom is access to their information granted in a decentralized environment?

While work like OAuth clearly has made progress in this area, the ability for users to define groups and sub-groups, with different degrees of trust, is needed.

Lastly, there is the question of trust and privacy of the data itself. In a decentralized environment where data is integrated across boundaries, the ability to trust data itself, not just the identity of other people, is crucial.

Distributed Architectures and Business Models

Social Networks have indeed a business potential but cannot be compared to traditional markets as is. Considering Social Networks' participants as mere potential costumers would be inappropriate; on the contrary they have to be regarded as active stakeholders of an ecosystem. The relationship between social networks and traditional enterprises could be complicated and it's likely that that an enterprises' excessive encroaching attitude towards them could be even harmful. It would certainly rise the subscribers' feeling of being manipulated hence reducing the attractiveness of a "private" ecosystem, which is the key success factor. Social Networks have to be distributed but still open, the traditional costumer exclusive "ownership" approach pertains to legacy paradigms. This session should identify the required actions for rising Social Networks potential and provide instruments to quantify their possible role in emerging business models.

There might be several monetization approaches. A Social Network has a value "per se" and could be sold to third party service providers; it could be served to an existing network of costumers as a free commodity; or be a new stand-alone economic ecosystem in which subscribers have an active role in the value chain; or even a mix of these scenarios.

Deeper and Adaptive User Experiences

The social network platform or agent can detect the type of device on which the user is connected at any point in time. When there is high speed, low cost network bandwidth available, this can be used for rich and immersive experiences. If the user is handicapped, the platform can adapt to provide compensatory interfaces. In the mobile handset, the use of widgets, the integration with the camera phone, the phone's addressbook are all ways to help users move seamlessly between their physical and digital words which do and will continue to co-exist. In this session we will discuss how intelligence in devices and networks can help users feel more connected with their communities and community sponsors.

Context and Communities

Contextual awareness adds value to consumer social networking services. Community operators can extract and leverage users' location information to provide more appropriate recommendations and to increase the satisfaction of social experiences between community participants. Unfortunately, every platform which uses context data today is developing technology and systems independently of other context management platforms and of the community messaging and advertising platforms to which these contextual data producers are complementary.

In the past, mobile network operators controlled all contextual data on subscribers. But, recently, with GPS in devices, services have begun to capture and leverage the location information gathered and/or managed by another service. For example,one application might determine a user's location and store it on a server for other applications to access and use.

In this session we will tackle multiple problems frequently associated with context sharing within and between communities and context service platforms.

Venue

See the venue page.