As promised below is part two of my mini-reviews of papers submitted to the W3C Workshop on the Future of Social Networking, including the three late papers, an interesting related paper by google (pdf), and Danbri’s take on Foaf in 2009. The workshop starts tomorrow. Part one of my reviews (papers 1-42) is here.
Most interesting to me: 43, 45, 50, 56, 57, 60, 61, 63, 64, 67, 68, and especially 52 which raises a lot of important points about what can or cannot be done with your harvested data (if anything).
Themes: there are an awful lot, and the program committee have done a good job in turning such a bunch of disparate material into a agenda and set of discussion points.
My take on the main themes from the papers:
- data silos problem and solutions; portability of data, policies and permissions
- trust, authentication and permissions
- semantic activity streams
- ownership of data created by networks; what can be done with it; data mining; creative commons for personal data
- identity across sites; mobile operators as brokers
- location awareness, apis or markup;
- context awareness, sensors and apis or markup for these
- accessibility and web 2.0
- business models
- best practices documentation
foaf, oauth, openId, sioc, Dataportability, hcard, vcard, atompub, xdi, NFC, Doap, opo and similar, openDD, OMB … and many more
Reviews 43 – 72
43. FOAF & SSL: creating a global decentralised authentication protocol Henry Story, SUN
Protecting rdf resources using foaf and ssl. Idea is that the user can identify themselves using an ssl certificate in their browser which refers to their dereferencable id (the #me in their foaf file) which means that the public key in the foaf file can be checked against the one in the certificate, and then access granted or not depending on some friend-related or other algorthm. Interesting, and has several implementations (what could be the relationship with openID and oauth, if any? are they all complementary?).
44. Managing Social Communications Identities (pdf) Óscar M. Solá, Telefónica I+D
Insteresting idea about linking users’ social and communication identifies in a secure and private and configurable way by a ‘social broker’. The idea being that you don’t have to know the phone number of a person, or their email, in rer to be able to contact them (provided that they have specified that you can contact them).
45. Current issues with Social Network Representations (pdf) Peter Mika, Yahoo! Research
Describes a view of a company getting to grips with using semantic markup: and the phases of microformats, rdfa; the need for mapping between Foaf and Vcard, lack of best practices for some types of rdf vocab mixing. Argues that vocabs should be produced using existing data about what people are willing to expose. Argues also that aspects of rdf are too hard to grasp or communicate. Emphasis on agreements on how to use existing things rather than creating more formal standards. Interesting to see a commercial point of view in this area
46. Social Networking Segmentation: Celebrating Community Diversity in a Framework (pdf) Christine Perey, PEREY Research & Consulting
Characterises different kinds of communities offering different sorts of experiences or services for mobile and static devices (and things in between). There are two classification systems: why the user is there (professional reasons, entertainment) and complexity of features (which are often related to whether the network is aimed at mobile or static devices). Argues that a widely used classification system would allow networks to “comunicate with their target market segments” and differentiate themselves more quickly.
47. It’s all around the domain ontologies – Ten benefits of a Subject-centric Information Architecture for the future of Social Networking (pdf) Lutz Maicher, Benjamin Bock, Topic Maps Lab at University of Leipzig, Germany
Argues that developing social sites starting with domain ontologies with object identity in topicmaps or RDF makes development easier in multiple ways (e.g. ontological flexibility in development, easier to localise, identity awareness.
48. Social Networking: Power to the People (pdf) Stefano Bortoli, Paolo Bouquet,Themis Palpanas, University of Trento, Italy
Argues that users should own their data and be able to move it around rather than being locked in to a particular social network. Argues that foaf is not sufficient for the needs of a decentralised network because: doesn’t make enough distinction between types of relationships; is public to all; doesn’t provide a solution to identifying people and other things uniquely. They are building tools under the OKKAM EU project, e.g. http://www.foaf-o-matic.org and a distributed system for generating and storing unique identifiers on the web (hm, what’s wrong with URLs?)
49: A Telecom Italia view on the future of Social networking (pdf) Claudio Venezia, Telecom Italia
Clear statement of the company’s interest in areas of standardisation or endorsement that W3C could undertake, in the araes of identity, portability, privacy, user experience; e.g. endorsing openID, foaf and sioc or similar standardisation and specialised URI schemes; endorse or create something like IdM or oauth; best practices for mobile user interfaces, and several more. Plus a plea to bear in mind that these networks need to be monetized. Worth a look.
50: Beyond Eyeballs: Improving Social Networking Metrics (pdf) Christine Perey, PEREY Research & Consulting
Argues that current metrics for evaluating social networks do not make for very interesting or useful analysis (page impresssions / month and new accounts). Suggests that a common framework would allow better allocation of resources and if shared would enable better comparison of sites. They suggest types of user: joiners, collectors, critics, creators; user profile metrics (e.g. ‘gardening’, ‘policing’; ‘giving’ and ‘receiving’ actions respectively examples could be rating and viewing others’ contributions) and various others (number of friends, various frequencies). Also 17 community metrics (e.g. user funnel, total number of pieces of content added per month, percentage of different users types). Interesting because these types of stats drive allocation of resources in many companies.
51. NewBay Position Paper on Mobile Social Networking Stephen Farrell, Bill de hOra, NewBay
A list of recommendations for w3c action is at the end, as “a software and services provider to mobile network operators”. Their view is that the issues with social networks on mobiles have to do with user interface issues (e.g. web 2.0 self-updating pages etc) and ‘irritation issues’ – a constant stream of events may be irritating in a mobile context when it is not in a static context. They suggest that brokerage by mobile operator may be the way forward to transmit preferences of the user. Not sure I completely understand why this is the best option though.
52. Social Networks as a Future Geographical Data Source (pdf) Ian Holt, Jennifer Green, Ordnance Survey of Great Britain
As a data vendor, the OS has been researching data mining in social networks (in this case to extract vernacular placenames as areas on a map). They are interested in the legal and intellectual property questions raised by this, the possibilty of standardising something like dataportability.org for data sharing standards; and whether such data is a marketable commodity, and what the need for anonymity is in these cases. Very very interesting questions indeed.
53. Open Platform for Multichannel Media Distribution Management Roberto García, Juan Manuel Gimeno, Universitat de Lleida, Spain
Describes an EU project to create an open platform based on semantic web technologies for the distribution of content from small and medium content providers. It will have digital rights management features based on a copyright ontology and will use user tracking rather than DRM. Not immediately clear to me how this is relevant to the workshop.
54. Mobile Social Networking: Two Great Tastes John Kemp, Franklin Reynolds, Nokia
Describes various aspects of mobile phones that makes social metworks on mobile phones different to social networks on static devices on the web. Interest in radio capabilties like bluetooth, GPS; capacity to interact with the real world using 2D barcodes; and they’re always with us. Privacy implications of phone number as a unique identifier. Suggests a distributed architecture for social networking using the processing power of these devices and not dependent on an always-on connection. Argues this would need more interop between sites; doesn’t really explain why.
55. Social Networks in Life Sciences: Defining and Enabling Appropriate Roles to Create an Atmosphere of Trust and Security (pdf) Hans Constandt, Adrian Seccombe, Robert Sweet, Yijing Zhou, Susie Stephens, Eli Lilly
Interesting idea somewhat related to paper 52, about the possibility of using semantically enhanced data from social networks of individuals with a particular disease, and similar questons of anonymity and tracability of using of this sort of data.
56. Towards an OpenID-based solution to the Social Network Interoperability problem (pdf) Michele Mostarda, Davide Palmisano, Federico Zani, Simone Tripodi, Asemantics
The paper describes a piece of software: an implementation of OpenId that can have connectors that connect the user to various social networks, including via the open social API, and aggregate their data from their networks, filtered if they like to different personas, on to one or more personal pages. the ppaer also talks abot a generalisation of this approach, termed the “Global Social Platform”. Interesting; a bit unclear to me whether the current system requires youy to give away your passwords or not though.
57. Collaborative Filtering and Social Capital Peter Ferne, Jiva Technology
Interesting summary of some aspects of social captial (‘whuffie’), including recommending people. Discussion of the complexity of measuring social capital. Idea that trustworthy systems systems require openness. Good set of links to follow.
58. Applying an XML Warehouse to Social Network Analysis (pdf) Benjamin Nguyen (University of Versailles), Antoine Vion (University of Aix-Marseille II), François-Xavier Dudouet (Université Paris-Dauphine), Loïc Saint-Ghislain (Ecole des Mines de Nancy)
Describes a project to analyse data from the W3C mailing lists, using XML databases and XQuery. They have used this to create networks via co-authors as well as other types of analysis. Interesting work, but not perhaps relevant as it stands to the workshop?
59. Mobile Eco-System: The Need for a Mobile Markup Language Nicolas Belloni, Mattias Rost, Future Applications Lab/Mobile Life Centre, Stockholm, Sweden
Argues for the need for a markup language for mobile services, for “absolute location, sensors, near-field communication, proximity of other users or services” to improve access for creative to this information. They are developing prototypes. Formatting the text would have been nice
60. Ten Theses on the Future of Social Networking Harry Halpin, University of Edinburgh
Paper describes the elements required for opening up data silos, arguing that the technologies are there already, and what’s needed is openID, oath and foaf; He argues that it’s in the interest of producers and consumers of data to have consistently structured data. He emphasises that apis should not preclude what is being used now, and that we should use data about what is being used as a basis for standardisation. He wants to use rdf, RIF and the W3C, cooperating with dataportability.org to void duplication; and to include provenance. Suggests a best practice recommendation. I’ve a lot of sympathy with his arguments – I’d like to see a sample implementation.
61. The Relationship Layer and the Secretary (pdf) Dewey Gaedcke, Minggl.com
Interesting and clearly argued short paper about how a secretary-like application which could prioritise and deprioritise and reroute information in a similar way to humans, by using statistics about how we interact with our peers. Argues that a minimum set of things needed are: global identity for a user and mapping to app-specific identities; open api and semantic event-type data ‘actionstory’.
62. Mobile Video Improvements to Enhance Mobile Social Networks (pdf) Tim Hyland, Dwipal Desai, YouTube
Argues that it’s important for social networking on mobiles that a consistent way of inline video playback that works on all handsets is decided on – doesn’t matter if it’s html5 video tag, flash or something else – but it is important, and users will expect it, as it’s so often used in static social networking.
63. Social Media in eGovernment John Sheridan (The [UK] National Archives), Kevin Novak (The American Institute of Architects), José M. Alonso (W3C/CTIC)
Explores some of the implications of government interaction in social networks. The ‘OS’ question again pops up – who owns the data created by these networks, can it be usedd for anything else, and how can it be anonymised if so – and what are the privacy implications? Interesting read; it’s come out of discussion at the W3C eGovernment Interest Group.
64. SIOC: Content Exchange and Semantic Interoperability Between Social Networks John G. Breslin (National University of Ireland, Galway), Uldis Bojārs (National University of Ireland, Galway), Alexandre Passant (National University of Ireland, Galway), Sergio Fernández (Fundación CTIC), Stefan Decker (National University of Ireland, Galway)
A paper describing the features of SIOC and how it interoperates with other onotologies, enhances site interoperability, and is used in multiple tools. SIOC describes idems at the level of containers and content items – blog, blogpost, items, bookmarks, comments. In the furure would like to get closer integration with OPO (online presence ontology, paper 12). Argues that for these reasons W3C efforts in this area should include SIOC.
65. Integrating Social Networks and Sensor Networks John G. Breslin, Stefan Decker, Manfred Hauswirth, Gearoid Hynes, Danh Le Phuoc, Uldis Bojārs, Alexandre Passant, Axel Polleres, Cornelius Rabsch, Vinny Reynolds, National University of Ireland, Galway
The 10(!) authors provide some usecases for sensors and social networks, and suggest that sensors can create semantic data about a user’s activities, and that they can extend and create social networks. They think that “some interaction between the Semantic Web and the Mobile community within a W3C group could be beneficial to this convergence”
66. Enabling Trust and Privacy on the Social Web Alexandre Passant (National University of Ireland, Galway), Philipp Kärger (L3S Research Center, Hannover, Germany), Michael Hausenblas (National University of Ireland, Galway), Daniel Olmedilla (Telefonica R&D, Madrid), Axel Polleres (National University of Ireland, Galway), Stefan Decker (National University of Ireland, Galway)
A discussion of trust and privacy and the relationship to the semantic web stack; they believe semantic web techniques could be used successfully for trust and privacy, for example to share photos of multiple sites to a small group. They are interested in policy-based approaches, and agreed models for defining policies and authoritativeness.
67. The Tangled Web We Weave (pdf) Greg Howard, Rajesh Kuppuswamy, Kaushik Sethuraman, Microsoft Corporation
Paper arguing that accessing social networks by mobile devices will require techiques either for agfregating multipel networks in a single UI or a way of quickly flipping between networks – and cruically matching up friends over networks for the user, on the device itself. This could be done in an automated way to suggest links, and also proivde a way to manually link them. Interesting.
68. Business Context Impacts on Social Networking Mary Ellen Zurko, Werner Geyer, IBM
Interesting again. Lotus describes two of its products that allow companies to manage their business contacts. They are interested in using openID to allow trusted partners to access this information, and perhaps open social. They want to be able to define different types of buiness relationships.
69. A Vision of an open Platform – The Enablers Perspective (pdf) Bastian Pfister, Roman Hänsler, aka-aki
A small, mobile social hetworking company wants to make a social networking tool that would enable users who meet in physical space to interact with each other on social networks. Requires social networks not to be data silos. Argues that consumers will expect this sort fo thing to work and will be surprised when it doesn’t. Additonally data costs need to fall.
70. Rethinking digital object, Rethinking information relevance (pdf) Yuk Hui, Centre for Cultural Studies/ Department of Computing, Goldsmiths, University of London
A highly philosphical short paper. He wants to investigate how relevant data is to an individual – in order to enhance ‘ambient findability’. The work is at an apparantly early stage and is not (yet anyway) relevant to the workshop.
71. DMM: Digital Me Management Karl Dubost, Olivier Théreaux
Outlines various issues about data silos, data ownership in these silos and various access control approaches. The authors’ position is that each of these need to be addressed.
72. Different groups with share agendas (pdf) Elias Bizannes
The author (who is vice-chair of dataportability group) outlines the groups he is involved with (open social open web foundation, DiSo), and states that they are not competing but offer overlapping and complementary benefits. The data portability project has spent 2008 establishing governance and process, and various things are now in progress: a tool to assess sites and companies with respect to open standards dataportability supports; a creative commons for personal information; a healthcare taskforce.
Semantic enhancements for social networks Rigo Wenning, Ivan Herman, W3C
Argues that it is not simply enough to make data silos permeable but there’s also a need for the ability to move policies, reputation, traceability and privacy and access controls between networks.
Web 2.0 and the Visually Impaired Learners (pdf) Nantanoot Suwannawut
Describes various common problems with web 2.0 and accesssibility notably ajax updating bits of the page which screen readers are unable to track; some sites do not alow assistive devices to be connected; similarly there are no guarantees for new devices. Text alternatives are often not available for video content. Capchas are not usable. Large parts of the web are no longer accessible and so some groups risk exclusion.
Ubiquitous, social networks in the street (pdf) Marc Pous, Luigi Ceccaroni, Manel Palau, Victor Codina, TMT Factory
Describes a service that suggests personalised activities, people to do them with, and how to get there, using recommendations, sensors, social networks, time and geolocation. Needs objects, locations, people, events etc to be described in interoperable and portable ways.
(Under)mining Privacy in Social Networks (pdf) is worth a look – google on privacy issues with social networks – unexpected events in activity streams, accidental linking of personae, and datamining for merging. And using the social graph for mitigating these.
Finally, last but not least: Danbri’s: foaf in 2009 plan outlining goals such as: evaluate effects of people not making their own homepage by hand any more; best practice for sites which expose foaf; impact on regular users of aggregators; creative commons for personal data. Plus technical issues: vcard and portable contacts and foaf; foaf and openID, oauth, atompub, webdav, ssl, pgp; trust and provenance; crawler stats and rest apis to large aggregators; searchmonkey, google social graph. Aiming to have regular meetings (f2f where feasible and regular online meetings) and a decision process / calendar for the core vocab.
Hope it all goes well!