We introduce at a high level our approach to system design and the technical implementation, leaving the details to a separate section, System Requirements for Sigala.

We start by proposing a few basic principles by which we aim to realize Sigala as a properly working system, i.e. to ensure that we are building the system right as well as to build the right system:

  • Fit for purpose: From a user perspective, the first encounter is crucial to longer-term adoption. Many aspects make up the evaluation, but much rests on whether it ‘feels’ right, has something unique, allows further exploration without undue limitations. Once we start using it more than once, then we are showing some acceptance and commitment, i.e. we are trusting that it serves our needs appropriately.
    So we need to ensure that the design is fundamentally correct.
  • Simple and malleable: By their very nature, once SNS gain popularity, they can grow very rapidly. We are developing a subtle system, with potentially great complexity. We must start small, be agile in development and strive to keep it as simple as possible. The sculptor, Henry Moore, used to work on his designs in miniature, producing plaster maquettes1. They could be held in the palm of his hand and he was able to change their basic structure very easily and always retain an overview. Once he was satisfied, he then got a team to realize the full-size version, in bronze and other materials.
    Similarly, we can develop initially just a small skeletal model – think of it as a digital wireframe – to illustrate the basic shape of the system and once it has the right qualities, we may add to it – just as we would add organs, muscles and flesh to a human body.
  • Available to the right people at the right moment. The right people should be encouraged to join at the right moment. In particular, the timing of software releases (and its associated publicity), especially for the public, is a critical factor. Unlike a human body, a software skeleton may already provide a useful service before it becomes a fully-fledged full-scale system. We aim to strike a balance between premature release and perfectionism.
  • Developer Community. There will be no SNS without the people who write the software. So it’s essential to foster a skilled and enthusiastic developer community that is capable of developing the right system in the right way. As the Internet is fundamentally open then the developer community should be too and it will need to be continually nurtured.

    This merits its own treatment and is not covered in detail here, but we expect the views of open source advocates such as Eric Raymond and many others to feature prominently.

Architecture

We consider together the aspects of data and access (navigation) because they have aYin-Yang relationship and should support each other. In conceiving the data structures we have to think about how that data can be accessed by degrees, in stages, naturally.

Much effort has been made to develop the semantic Web; to make systems intelligent insofar as they can help establish the correct context of the given data. As this context is ever-changing as we search and browse, the semantics needs to change correspondingly. For software to be intelligent about the data it provides depends on data being structured in such a way that supports intelligent access and especially how people will navigate their way through the data.

Just as with navigation of the physical world, we need to store and present data at different levels of granularity. Hence the notion of semantic zoom should be a core part of the design, so that at successively deeper levels the nature of data elaboration depends upon the [semantic] context and can vary from level to level.

We plan for the network architecture requirements relating to the deployment as an Internet-based service. The specifics are to be determined, but the basic requirements include:

  • Scalable for the user – the service must work well for any number from a handful to a few thousand connections
  • Scalable System – as a service it must easily cater for any number of users, from a handful to millions
  • Sustainable – it needs to be easy to maintain and develop further

Client Support

This SNS should be easy to use anywhere.

  • Cross-platform, designed for emerging technologies, especially personally owned handheld devices such as smartphones and tablet computers as well as tablets, laptops, and desktops
  • Aware of latest interfaces, especially multi-touch.

User Services

The technical specification of user services should cover:

  • Communication – is all about messages of various kinds in various formats – extended content (blogs) and micro content (short messages). Standard protocols should be supported, but new ones also considered
  • Aggregation – to present more nuanced newsfeeds and other content, particularly based on different relationship types
  • Micro-transactions: based on the specific relationship types, a polymorphic service that supports many kinds of [small] payments, in terms of monetary payments, virtual goods, and physical goods
  • Localised Calendaring and Events – standard facilities for arranging gatherings will be enhanced by being culturally aware – in terms of special days, times, venues etc.
  • Sharing Narratives: some of the most substantial content comes from the sharing of edifying life stories structured through various kinds of narrative forms. Sigala should enhance this through its relationship structure.
  • E-Learning support: e-learning is a more formal complement to life experiences, and as such it should fit around especially the teacher-student relationship.

User Interface

The user interface needs to be well designed in order to gain broad acceptance, both in terms of how the site is presented and also in terms of how users contribute content. At the outset we should ask fundamental questions such as: who gets to see what and how should that be presented? Should everyone have the same interface or should they be able to personalise it?

It has usually been the case that SNS present just a small selection of views in a standard look and feel (layout, colour scheme etc.) – one for the owner of their space and the other for visitors. This is the case for Facebook. One exception to this was MySpace, which at the outset allowed users to determine some of the more visual elements for visitors.

The paradigm that provides the basis for views in Sigala is that of giving and receiving: the interface is the means of connecting the giver with the receiver. If personalisation is to be properly utilised, then the interface should be tuned accordingly and aim to facilitate that communication in an optimal way. Accessibility should be seen as a natural consequence of such a requirement.

Entry Dialogue: Identity and Identification

Entering and becoming a member of a social networking service is a substantial commitment – it could involve a lot of time and mental energy. So it’s not something that should be rushed into. SNS themselves can take some responsibility by providing a meaningful registration process. In Sigala we propose that registration is primarily about establishing identity and identification in a way that supports meaningful diversity, as described earlier.

When someone comes to register their identity is defined by entering some personal details. Some of these details are formal and factual such as nationality, the kinds of objective verifiable data that might be stored in an identity card. On the other hand some facts – such as ethnicity – that may be of defining personal significance might not be stored. Furthermore, other aspects such as religious belief are unlikely to be recorded, but they may be what an individual regards subjectively as important characteristics. If we are considering a global system then religion and faith have great importance because they offer immediate significant identification, i.e. two people of the same faith have some meaningful connection.

The system can encourage a newcomer to reflect on their identity – both what they regard personally, in private, and how they wish to present themselves to the world. Normally this is done perfunctorily in SNS by binary checkboxes (show/hide) with no guidance. Furthermore, data is presented with the same layout, offering the same priorities. Sigala cannot be expected to offer much granularity, but it can engage the newcomer in making basic meaningful choices that have an impact on how the service is delivered.

The process envisaged will start by inviting registrants to indicate what is important in how they identify themselves – not just in what they put in their summary profiles, but how they view the world2. Alongside, they should indicate levels of sharing of such information – from completely private to fully public.

The system then will then prepare views. For a basic system, this will be in terms of what profile data to show, but for a more advanced implementation the profile will also determine how to form the layout (for themselves and for others). For instance, someone who is ostensibly religious might have religious content and contacts more visible. On the other hand, for someone who regards themselves as religious, but doesn’t want to show this, recipients will not receive this impression from the interface, but may do so in other directed communications. Where both the giver and receiver indicate common elements in their identity, these commonalities may be highlighted subtly. For people who are connected according to a particular relationship type, this can be reflected in providing different profiles (including portrait images etc) in the context of, for example, shared identity3. This could be refined still further to project other aspects – for instance an old school photograph in the context of school reunions.

Individuals should be able to preview how their content will appear to others, a facility that has already been made available in some SNS such as Google+. This will indicate the range of views, which may be based on each direction4. The subsequent interactions in each of the directions then serve to establish the identity; in this way identity is affirmed as multi-faceted and as such it becomes diversified. Technically we may regard a persona as a subset of these facets that reflects naturally the given relationship context.

In this way, identity really becomes operative in the SNS and contributes to the system’s personalised architecture in a way that is affirming and enriching for the participant.

Making a connection

Having registered on the service and familiarized oneself with it, the next step is to make connections.

If we travel abroad it is customary to observe local traditions. For example, in Thailand, we greet each other with a ‘wai’ (with hands together, palm to palm and a bow of the head). The system can attempt to reflect this in a simple yet faithful way based on the preferences registered, particularly nationality and locality. Every contact is an opportunity to learn with a light touch – it should not be imposed heavy-handedly.

Site Presentation

Mainstream use of handheld devices with small screens has led to mobile-first design, making it increasingly important to design interfaces that are easy to understand by the vast majority of users – both the individual pages and the navigation between them. This generally requires an uncluttered view, so the choices presented at any stage need to be carefully evaluated.

If a site view is personalized, then how are views shared when making contact with another person? In the physical world, the shared space (home, café, etc. is the same), yet how they are perceived may vary considerably. When an individual acts autonomously, then the view of the virtual world naturally should reflect the preferences of this person. However, when involved in some interaction then we should allow the site presentation to be combined with the preferences of the recipient, which suggests some kind of filtering according to the specific relationship type. Generally, the system should promote quality over quantity where a conversation has much greater value than pressing a ‘like’ button.

Personalisation

Personalisation may start conventionally with customisation options for look and feel –at least for use of the SNS there should be options that suit different age groups (e.g. young children’s interface). However, personalisation in Sigala should extend the functionality of a service, fine-tuning it to meet individual and collective needs. Further, a service that supports personal and work activities should have options to clearly differentiate the various kinds of activities to maintain separation of concerns.

The facilities should enhance access and protection at the same time. Hence it should be easy to find people and knowledge, whilst making it straightforward to check what one is sharing with whom. The principal of not taking that which is not given should also be applied so that there is always opt-in (changes should not be slipped in without the user knowing about them; not doing this has cost Facebook legal action5). A succession of these can lead to a breakdown in trust. Furthermore, the system should ensure that personal data is not propagated without intentional consent of the person. This should guard against the case where Joe contact shares some personal information with Jane that ends up with a 3rd party service, e.g. an app manufacturer, without Joe’s knowledge.

To help facilitate connections, the user can define their own relationship types to extend the universal and localised relationships. As with these standard relationship types, user-definable relationships will require mutuality to be respected. Mechanisms need to be provided to facilitate their creation, perhaps in group activities.

User Interface Design

To make the site attractive we shall use the following elements:

  • A relationship ‘surface’ or ‘tableau’ offering the flexibility of a work surface –starting clean and empty, onto which you can add what you want.
  • Full touch screen support operating in multiple modes that variously allow focus on the task in hand or movement between people and views.
  • Abstraction and local contextualisationAs the social space is vast, the interface design should avail itself of research into navigation of spaces that cannot fit onto a fixed-sized display. This research has been actively explored for handheld and mobile devices out of necessity.

    It’s noticeable that the ubiquity of mobile devices with a wide range of screen sizes has prompted considerable effort to redesign user interfaces with special attention to abstraction. This has been particularly manifest in the Microsoft Metro interface design that carried over from Windows Mobile to Windows 8, thus a reversal of earlier design paths where the opposite happened for Windows CE in the previously niche market of handheld devices. The attention given to abstraction is very significant as it necessitates careful consideration of the psychological implications – how the mind handles switching between different views. Microsoft, realizing they made a misjudgement among desktop users, subsequently launched Windows 10 as a mixture of the paradigms.

    At the same time, there is extensive research into the use of special lenses, especially peephole methods that retain a depiction of the whole whilst navigating particular sections under magnification. Some of this research has particularly demonstrated its potential for navigating 3D spaces6

    How effective this can be will depend upon the data architecture. As mentioned above, we plan to design the data to support semantic zoom and for its implementation we propose a novel approach navigation that respects the multidimensional nature of relationships, with non-linguistic visual aids for accessibility. Furthermore, we may enhance the interface using light aesthetics, making the best use of the fact that computer displays are backlit; natural effects – light and shade, auras, radiance, glows, dimmer effects, gradual brightening etc. The default colour scheme for Sigala should uplift, even aspire to a celestial aura.

Multiple Input Modes

Whilst it is still the case that most communication is terminal-type text-based, albeit with some embellishments and multimedia adornments (photos, audio and video), the range of user inputs continue to grow and become more naturalistic. With the growth of touch screen devices and more recently contact-free devices, instant messaging should emerge beyond its textual confines, to include graphical forms of communication as described in the proposal for graphical and symbolic instant messaging.

Looking to the future, discussions on HCI are moving from ergonomics around a fixed workspace into areas that make much fuller use of the senses, especially kin[a]esthetics, which take account of whole body movements and physical gestures (consider e.g. greetings such as the Thai bow with hands palm to palm – the Wai); one branch haptics, concerns the sense of touch and how feedback may be gained that way, offering new methods to aid accessibility. These emerging areas for interface design are already quite in evidence in game technology such as Microsoft’s Xbox Kinect, which offer considerable educational potential7.

Security

Security needs to be rigorous in technical implementation, but also of great importance is the education of users.

The basic requirements briefly stated are:

  • Registration – needs to ensure that anyone can register, but in a way that prevents or minimises the chances of masquerading.
  • Authentication – access to accounts need to ensure a certain degree of security, but what should that mean in practice? Strong passwords? Tokens? Biometrics? 
  • Data security – passwords need to be stored as securely as possible using an architecture that has several levels of defence. Personal data relating to profiles also needs to be stored as securely as possible.
  • Authorisation – users should only access what they are entitled to. This must hold especially when data is propagated, for example content syndicated to people of a particular relationship type.
  • Some basic education in security and the safeguarding of privacy – users should always be careful about divulging personal details. The new relationship model should help to make this easier, but guidance should still be provided8.

    The system itself should help protect against phishing and limit the potential damage of worms and other contagious malware.
  • In support of the previous item, Sigala may observe a policy of not allowing logins through third party sites; Sigala should not provide ‘friend finder’ services that invite the user to enter credentials belonging to a third party service.
  • Similarly, protection against social engineering may include tests against social bots that impersonate real people to steal personal data – “The researchers found that the more friendships people had on Facebook, the more likely they were to accept the ‘fake’ friend.”9.

Among other significant factors that impinge on security and privacy are commercial access, particularly to personal profiles. Concerns have been expressed over the years by security experts, but it has taken many breaches to get to the point where this is recognised as a major public issue10. Potential problems can be mitigated by first identifying commercial users, making them explicit to users, ensuring that profiles delivered to commercial parties are anonymised, and that there is greater use of aggregate data.

Personal Safety

Personal safety is one facet that becomes prominent with increased exposure – to other people and the content that they make available. The flip side of ‘affordances’ that create opportunities for learning and creativity is an array of misdeeds that can deeply harm society. This places an obligation on system designs to be robust. The IETF netiquette guidelines showed awareness of the potential minefield in online communications in the early 1990s when the services available were more limited. Given the exponential growth of participants, services and bandwidth these issues are magnified and deserve much fuller treatment – not least in basic research. We have to consider with some earnestness matters of social protection, which more than ever require qualities of vigilance and heedfulness.

Whilst shocking headlines are not a reason to close down a service per se, they may be symptomatic of something inherently wrong in system design. We list a selection of such headlines below to draw attention to salient issues and then suggest how we intend to address them.

Protecting against Malevolent People

It is said jokingly that it takes a human to err, but computers to make a total mess. A tragic consequence of this fact plus the prevailing openness of networks means that criminals and other malevolent people can take advantage of this to multiply their activities11.

The recent nature of mass public participation and the delay in the reporting of figures means that academic literature is still quite sparse. In 2010 a paper by Mitchell et al looked at established cases of crimes against minors using data from 200612. They found that whilst the number of cases established has not been large, SNS-related crimes were even for the period in question (before the rapid growth of Facebook) a relatively high proportion. The paper indicates (p. 186) the variety of ways in which criminal activity can be undertaken; indeed the SNS support for proliferation is a topic of current concern. We note especially in the conclusions that “prevention messages should target youth behaviors rather than specific online locations.”

Much more recent data has been used in a 2011 study that was carried out into cyberstalking; in their research, a team at the University of Bedfordshire conducted a survey over 6 months between September 2010 and March 2011. The 353 responses revealed that the exposure to bullying was a major phenomenon and SNS like Facebook were the primary vehicles13. In response, the authors of the report recommended that sites signup to a code of practice setting out how they will deal with threats and abuse, including clear guidelines on the reporting process. These recommendations will need to be included in the requirements for Sigala.

The potential problems are exacerbated as young children join SNS, often under pressure from peers – at that age there is not so much awareness of privacy settings and the effects of bullying are deeper14. Sigala can protect children by identifying the parent (or guardian) – child relationship type as a dependency and in such a relationship act as a gatekeeper or moderator. One function of such a role can be right at the beginning when the guardian registers a minor so as to ensure the right data and permissions are entered. In their ongoing role as guardian they can be kept informed in various ways of how their child is using the network – for example, if a young child wishes to connect with someone who is not connected to their guardian or a relative, then the guardian receives notification and might be expected to give permission.

The system can make further background checks on that person to try to validate the details that they have entered, both quantitative and qualitative: for example, checks can be made against authoritative sources of information such as census data, and an analysis carried out on their network, perhaps with input from members of the network acting as referees. If there are question marks as a result of such analysis, then warnings may be issued with varying levels of severity, as with anti-virus software. Some aspects might be implemented as a scoring system and the level of control should probably depend on the localisation to reflect variation among cultures, but great care should be taken to prevent the system from being exploited so as to deny human rights. 

Protecting Property

Similarly, a risk of burglary has been identified by sharing too much in the status updates15. If privacy settings are too open, then unwelcome visitors may take advantage of knowledge about prolonged absences.

Once again we assert that the homogeneity in connection, so that everyone is linked with each other in the same way, combine with the openness of the network to create vulnerabilities.

Inappropriate content: Protection, Monitoring, Enforcement and Cleaning up

All SNS need to protect their users from inappropriate – especially illegal – content. The severity of this issue will depend to a major extent on how the system is designed – prevention is better than cure. In a system where ties are weak, then access is made easier for perpetrators. In the absence of proper guidance there is also a greater tendency to make public accusations on the spur of the moment and risk legal transgressions16. This is why some constraints are needed in the network design – as provided by the separation of concerns – and why guidance is needed, as discussed above.

Software systems, particularly those that deploy A.I. are able to provide sophisticated system monitoring, but it can never be 100% accurate – there will always be serious cases that elude detection and false positives over legitimate activity.  The eyes and ears of people in the various networks remains vitally important to maintain vigilance.

When it comes to enforcement, who should do the policing? How can one know that they are suitably qualified and trustworthy? For large systems it has been a major operation, one that is generally kept quiet, somewhat ironic for networks that emphasize openness17. When considering our entire society as a social network, we know that we have within this society appropriate means for people to assume official roles – people can be recommended through their personal networks and we may have, for example, relatives and/or friends who we know are in such roles. In other words, the social network finds within itself the means to police itself and this process is open.

A general term for regulating content in SNS and online communities in general is moderation. The fact that internal moderation is inadequate is evidenced by the proliferation of commercial services that provide forms of external moderation, both automated and human. These are typically used by and for organisations to protect their brand and reputation, but also to watch over minors, particularly to protect personal data. They can be effective for those who pay for their services and have been a source for useful guidelines, but such guidance is not much in evidence with regard to the likes of Facebook. For the majority of minors, the risks remain considerable.

There are actually some large systems where moderation is conducted entirely internally, a notable social media example being Wikipedia. A well-designed system will be sustainable in terms of moderation, but this will be a task made easier if all the protective relationships (such as friends and relations) are properly involved. Whilst in some cases these are the sources of the problems, in general these are the exceptions to the rule and society is better protected where these structures are in place and supported and guidance (as to Sigala) is taught and pursued.

Safety Engineering Approaches for Human Protection

What is safety? We may define safety as freedom from harm. In Computer Science there is the field of safety-critical systems in which safety is commonly defined as protection from critical loss, principally of life. Along similar lines we may more generally define in society asa whole safety as protection from fatality and significant degradation in the quality of human life – not just physical, but emotional and mental as well. In the context of SNS, safety means in particular protection of one’s regard – one’s profile and reputation. It is therefore important to realize a system that is robust with respect to social protection.

Whatever the precise definition of safety, it is good practice to put in place a safety lifecycle whereby potential dangers (technically called hazards) may be identified, risks assessed and handled accordingly. Safety issues need to be treated as intrinsic and handled systematically. Further, not only should the design reduce risk and its effects, but software should be analysed for faults, how it can fail – traditionally handled in safety engineering by disciples such as fault tree analysis (FTA) and failure modes, effects and criticality analysis (FMECA). Such methods have traditionally been oriented to physical factors, so with appropriate modification they should become suitable for SNS, whence following these approaches should at least improve the safety for all users.

From the system safety point of view, we see that making things too easy has caused problems with individual safety in terms of one’s profile and reputation. From this perspective some of the most popular systems are severely lacking in robustness: when accepted connections have access everywhere it is like a ship that as soon as just one compartment floods, then the entire ship goes down18. Commonly, someone shares information about another person in a harmful way and because of the recursively open connectedness, it spreads very rapidly and causes great long-term damage to relationships that had been built over time on trust.

In response, Sigala offers protection in two dimensions – horizontal, in terms of the separation of concerns into different types of relationships; and vertical in terms of depth of relationships. It means that sharing with specific kinds of connections of at least a certain depth will help limit indiscriminate propagation.

At the intersection of people’s preferences

A major challenge is dealing with content involving more than one person where there is the intersection of people’s preferences. This is commonly brought up in tagging and sharing photos. One person may like to tag a photo from a social gathering, but not all the people tagged may want the photo to be widely shared.

The principle operating here can be “not taking that which is not given”. An individual can specify their sharing options in terms of each direction, e.g.“family photos in which I share only with family” should only be shared among family members. If a family member wants to share this photo more widely then they will be prompted to seek permission from the content creator.

With regards to tagging there should be due attention to privacy.

Hosting

The technical problem of hosting reliable scalable services has been solved with cloud services, as amply demonstrated by large enterprises such as Google and Amazon. None are perfect, but they are good enough to be regarded by most users as generally dependable; technically they should be fine for Sigala.

What is less clear, though, is the question of where sites should be hosted and the respective legal jurisdictions. Large companies from the West have experienced difficulties in delivering their services in parts of the world where the culture is markedly different.  On the other hand, these same companies have also taken advantage of generous tax arrangements in various countries.  Sigala should aspire to improve on their track record by virtue of observing ethical finance and taking better account of cultural variation, societal structures and so forth.

These are the main areas for consideration, which we cover in more detail, starting with system requirements.

Notes 

1 See e.g. Bourne Maquette studio, Interactive Tour of Perry Green, Henry Moore Foundation, http://www.henry-moore.org/pg/interactive-tours/virtual-perry-green/buildings/d

2 Compare with e.g. UK National Statistics office guide to Identity:
http://www.statistics.gov.uk/hub/people-places/people/identity (archived)

3 This is discussed by the Buddha in terms of states of consciousness in the Potthapada Sutta [MN 1]; the key term is atta patilabha, which translates approximately as ‘acquired personality’.

4 The preview function resembles the need for coordinators in e-learning systems to be able to ‘switch user’ to see the course views provided to tutors, lectures, students, and all the other key roles.

5 BBC News Business Facebook settles privacy case with US regulators. 29 November 2011.http://www.bbc.co.uk/news/business-15953414

6 See Ka Ping-Yee. 2003. Peephole Displays: Pen Interaction on Spatially Aware Handheld Computers, Proceedings of the SIGCHI conference on Human factors in computing systems, Ft. Lauderdale, Florida, USA April 5-10, 2003, pp. 1-8, ACMhttp://dl.acm.org/citation.cfm?id=642613

7 Hui-mei Justina Hsu, The Potential of Kinect in Education, International Journal of Information and Education Technology, Vol. 1, No. 5, December 2011 http://www.ijiet.org/papers/59-R025.pdf

8 See e.g. SOPHOS Facebook Security best practices – Profile Information
http://www.sophos.com/security/best-practice/facebook/profile-information.html

9 BBC News Technology. Socialbots used by researchers to ‘steal’ Facebook data. 2 November 2011. http://www.bbc.co.uk/news/technology-15553192

10 See e.g. Mark Ward. Bruce Schneier warns ‘profits killing personal privacy, BBC News, 12 October 2010. http://www.bbc.co.uk/news/technology-11524041

11 BBC News Cornwall. Facebook and Bebo child sex abuse postman jailed 24 September 2010. http://www.bbc.co.uk/news/uk-england-cornwall-11403984

12 Kimberly J. Mitchell, David Finkelhor, Lisa M. Jones, Janis Wolak. 2010. Use of Social Networking Sites in Online Sex Crimes Against Minors: An Examination of National Incidence and Means of Utilization Volume 47 Issue 2, Journal of Adolescent Health, August 2010, pp. 183-190.

13 Jim Read BBC News Beat. 11 July 2011. Social network sites ‘have duty’ to stop cyberstalking http://www.bbc.co.uk/newsbeat/14085766

14 BBC News [Radio 5 audio excerpt]. 25 May 2011. Mark Zuckerberg wants pre-teens to have access to Facebook http://www.bbc.co.uk/news/technology-13546902Note that this actually consists of 2 interviews with parents about 10/11 year olds using Facebook. (Zuckerberg’s determination eventually led to the launch of the controversial Facebook Messenger Kids in countries deemed more ‘receptive’ to the idea.)

15 Liz Roberts Facebook status updates are ‘burglary risk’ BBC Midlands Today. 23 December 2010.
http://www.bbc.co.uk/news/uk-england-birmingham-12062331

16 BBC News Wales. 24 October 2010. Solicitors in Cardiff take on Facebook defamation cases http://www.bbc.co.uk/news/uk-wales-11610510

17 Iain Hollingshead, and Emma Barnett, The dark side of Facebook. Daily Telegraph. 2 March 2012http://www.telegraph.co.uk/technology/facebook/9118778/The-dark-side-of-Facebook.html

18 Have designers of SNS properly realized just how powerful are the effects and how important it is to have safeguards to respect the sensitivities of human relationships?