Introduction

A goal of this thesis is to bridge established social science research about self-presentation and impression management, with a practical, developer-centric perspective on building Social Web applications. I seek to understand how and why contemporary social networking systems both help and hinder individuals' expressive needs, and what people do to work around technical and social constraints they encounter during day to day use. What we learn from this, we can use to design and optimise future social systems for individual and collective empowerment.

The society from which I write is at present enthralled with online social media. It is rare to encounter someone who does not have a profile on one or more of the major Social Network Sites, most of which are household names. These systems have shifted from the realm of techie early adopters to indispensable tools of daily life at a rapid pace, and continue to evolve. Yet when I speak to people about their social media use, I inevitably encounter grumbles or complaints. Some people are frustrated by unintuitive user interfaces; others feel trapped or pressured into using particular systems because everyone else seems to be there, or it's become the only way to get anything done. Yet others are hooked, distressed to find themselves whiling away hours by watching other peoples' lives go by, but unable to tear themselves away. Others are driven away from interactions they want or need by harassment and abuse.

I am far from a techno-dystopianist. I believe strongly in the Web as a force for good, as a means to communicate ideas and share experiences across the world. I've used the Web for almost my entire life; it has been an outlet for creativity, a forum to learn about myself through the experiences of others, a means to maintain relationships at a distance, a provider of remote serendipity and opportunities, and an invaluable asset to lean on when travelling the world with some dependence on the kindness of strangers.

I worry about the digital shadow of myself which corporations and governments have access to. I am sure it is thorough and accurate, and that they could use it for all sorts of mischief. I worry about being manipulated without realising, about being tracked, about being backed into a corner with nowhere to hide. I worry more about the countless people this is happening to who do not have my considerable privilege which stems from my country of birth, my stable upbringing, my education, and the colour of my skin. The society from which I write is also undergoing some political upheaval. Ordinary people seem to be turning on each other, blaming people who are different from them for their problems; there is division and anger in the air. Or maybe it just appears this way through the particular lens of the Web through which I am witnessing current events unfold. Maybe it is somewhere in between. Maybe we have always been at war with Eastasia. It's not as bad as some parts of the world, nor many historical events, but it's not good. And social media, ever present, is playing a role.

Beneath the gauzy promises of democratized access to sociality, meaning, fame, and reputation, the business practices of the social platforms in and through which we self-present draw us all into privatized corporate strategies of social sorting, identity management, and control [hearn17].

This thesis is not about politics, but it is about people. It is also about society and community, and how we interact with others near and far through the possibilities availed to us by the Web, and how we come to know and show ourselves in the process. For this reason I first look to the social sciences, and Erving Goffman gives me a place to start, with presentation of self. To catch these ideas up to the present day, I lean heavily on the work of danah boyd, whose contributions are foundational to subsequent work on online, as opposed to face-to-face, interactions. Seeing the insights to be found through ethnography - through talking to people and observing them - I proceed to learn a great deal from my various study participants.

I am a Web developer by trade, and so I inevitably fall back to trying to design and build software to help with social problems, and I am often surrounded by others with this outlook. Now, I can do so whilst mindful of the continuously turning wheels and shifting sands of society, aware on some level of every unique individual who might pass through a system. That is not to say I think it's possible to accommodate everyone, but I think with enough energy and consideration it is possible to build systems which do not do as much harm as the ones we have today.

There are many ways to approach improving society through the Web, and the one I have chosen is based on the idea of putting people's digital representations into their own hands. There are many ways of going about this as well, and now I take you from this high level painting of the state of the world to a suddenly specific and technical detail. My efforts are towards creating standard protocols for Social Web activities, which allow other developers to build systems which can interact in an open and defined way. The Web Science Framework [websci] expounds the cruciality of Web standards for the progression of the Web, as a process of social negotiation which yields tangible engineering outcomes. The Web itself has always been decentralised^dns, and now I take inspiration from its inventor Tim Berners-Lee, who actively advocates for re-decentralisation of the social layer that has been built on top. Developers must build systems which respect the people who use them, because in decentralised systems people have the freedom to take their business elsewhere. I believe this changes the power dynamic in favour of the previously disempowered non-technical individual.

The process of creating Web standards, which I engaged with through the World Wide Web Consortium (W3C), involves nitty gritty technical work, understanding obscure specifications, practices, and web lore, and endless pedantic arguments. It is infuriating, but perhaps just as important to work through as it is to carry out the ethnographies I mentioned previously. We, along with other initiatives and open source projects, are inching towards progress. We are preparing the technical foundations on which social improvement can be built, when the time comes. Or at least, I hope so.

Throughout this work I advocate for particular visions of systems and technologies that I believe to be our best chance, however it is important to note that such systems are not an end in themselves. The social changes that technologies enable are what matters. Thus I want to emphasise that while Web technologies are constantly evolving, and subject to rapidly shifting political, economic and technological landscapes, that peoples' needs and desires, rights and responsibilities, are also reconfigured by changing technology. I hope to capture this feedback loop, and the enormity of an interconnected online and offline world, through this thesis.

Fortunately I am not alone in this. My work is situated in the field of Web Science, a multidisciplinary domain with a focus on the relationships between people and the technology that connects them. I lean somewhat on the narrative from the sub-topic of Social Machines, a concept which acknowledges the intrinsic co-dependence between humans and technology.

Why decentralisation?

I did not set out to explore the decentralisation of the Web specifically. When I started work on this thesis, I had a blog, and a healthy skepticism of social media, but no experience with federated or self-hosted social networks, Web standards, or decentralisation protocols. I stumbled across this world when I was searching for ways to empower content creators - people whose livelihoods were tied to (centralised) media platforms like YouTube and DeviantArt.

This thesis does not evaluate decentralised Web technologies and find them to be the most promising solution to the problems I described in this introduction. In fact throughout, I touch on new challenges that are a direct result of decentralising ordinarily centralised technologies or systems. In some cases I even go so far as to present possible solutions. In my concluding chapter, I summarise the most serious of these challenges. Rather, I decided that the decentralised Web vision of the future of online interactions was viable enough to be worthy of study; the timely formation of the W3C Social Web Working Group, and the specific focus of Tim Berners-Lee's MIT research group were contributing factors to this decision.

Accordingly, I do not mean for my focus on decentralised Web technologies to be read as a bias towards this as a solution. The bias which does exist, which of course influenced my choice of potential future to explore, is towards individuals and against for-profit companies; towards information access and transparency, and against manipulation and surveillance. Instead, the latter half of this work is a preemptive examination of a phenomenon which we may see come to pass in the near future, in relation to my core focus around presentation of self.

In this thesis, I ask questions about online self-presentation in the past, present, and future. I start by seeking to ease our understanding of what has gone before, or how we can make sense of historical research in this space. The landscape is changing under our feet, so I take a look at what is happening right now that will impact future research and evolving comprehension. Then I ask what we can take from this to prepare for and shape the road ahead.

Conceptual framework

R1: How can we access the bigger picture when it comes to understanding the impact of networked publics on presentation of self?

There is a multitude of work from the past two decades about self-presentation on the Web, from personal homepages and blogs to modern Social Network Sites. Ethnographic studies from social sciences and psychology investigate the impact of networked publics on people's everyday lives. Social media analysis studies from computer sciences look at network effects and find patterns in how people connect and the data they publish. Whilst there is usually some overlap, the former efforts tend to focus on people and the latter on technology. It would perhaps be unreasonable to expect anything else, for example, for social scientists to convey a profound understanding of the underlying systems their subjects engage with at every stage in their study. However, we are talking about socio-technical systems, and very complex ones at that. Networked publics cannot be properly understood without positioning an individual study against the contemporary background of what is occurring both socially and technically. Due to the sheer enormity of this task, much research fails to do this.

In order to make this easier, we need a device which allows scholars to access and organise concepts relating to broader socio-technical systems, so that they can situate their more in-depth niche or specialised analyses on particular topics. This can be achieved by means of a conceptual framework which captures a hierarchy of concepts that are applicable to understanding self-presentation in networked publics.

I begin by examining key concepts from relevant literature from both social sciences and computer sciences, and summarising a cross-section of findings from smaller scale studies about online social spaces in chapter 2. This sets the background for a series of studies of my own, which are described in chapter 3, and which are designed to tease out further diverse and novel considerations about social behaviour on the Web.

There are multiple interacting dimensions which ought to be considered when observing or designing online social systems, and until now it has been hard to find a coherent way of organising these. Chapter 3 proceeds to attempt to answer this research question with one possiblity for a conceptual framework design, based on findings from existing literature and my own study results.

Contributions:

• C1a: a novel conceptual framework - the 5Cs of Digital Personhood - that offers a consistent and comprehensive set of concepts and terminology for understanding the affordances and limitiations of self-presentation in Web-based social systems from a user-centric perspective.
• C1b: an up to date survey of existing work, which relates studies of the Social Web back to pre-Web work on sociality;
• C1c: a survey of features offered by a specific set of contemporary technical systems when it comes to profile construction.

Changing dynamics

R2: How does self-presentation change depending on the power dynamics of the Social Web services they use?

Next, I acknowledge the changing times of the Social Web. I believe (and hope) we are on the verge of an important transition from a world in which our personal data is stored and harnessed by powerful third-parties at great (but often unseen) cost to individuals, to the proliferation of technologies which enable people to be discerning about their choices of communication system. One route for this transition is by way of decentralisation, that is, by dispersing power from the few to the many. In terms of software for the Social Web, decentralisation entails making it possible for diverse systems to communicate without prior arrangement, to form spontaneous connections and pass data around seamlessly. Achieving this comes with both technical and social challenges, and to succeed would impact online social behaviour in ways we may not yet be prepared for.

The studies in chapter 3 contribute towards answering this question. On top of that, there is already much technical work towards decentralising the Social Web; in chapter 4 I describe a specific subset of this, and in chapter 5 I provide my own contributions to the field. Being directly involved in the technical work provides critical insight into the finer details of this question, from the 'other side' as it were, or the perspective of system creators rather than users who are at the heart of chapters 2 and 3.

Contributions:

• C2a: a set of critical dimensions to consider when studying identity performance through creative media sharing;
• C2b: a description of people's habits and reactions to different kinds of deception on social media;
• C2c: an analysis of attitudes towards self-presentation by people who control the technology behind their social media presence;
• C2d: a critical look at the technical directions taken by the W3C Social Web Working group, and the social dynamics of group participants which underly them.

Impact on practice

R3: What can developers do to adapt to or accommodate self-presentation needs of individuals?

Finally, I want to offer something of use to the designers and developers of future social systems. I seed some answers to this question in chapter 5 in an exploratory manner which is neither exhaustive nor conclusive. Nonetheless, my hope is that this prompts future work, theoretical and practical, around novel ways for the Social Web to empower its participants.

Contributions:

• C3a: a primer for the technologies produced by the Social Web Working Group, and technical guidance on how to fit the different specifications together;
• C3b: a prototype implementation of a personal social datastore, and a report on the personal impact of long-term use;
• C3c: a speculative design for a novel system for indirect communication between a profile owner and their audience, used for learning and meeting audience expectations in the moment a profile is viewed.

This thesis is roughly in two parts, and each part comprises a background chapter and a chapter containing my novel contribution. Part one, chapters 2 (background) and 3 (new work), is about the past and present of online self-presentation. Part two, chapters 4 (background) and 5 (new work) are about the present and future of online self-presentation should the Social Web become decentralised. The first half has the feel of work from the social sciences domain, and the latter half is highly technical. These two parts in combination are necessary for providing both a novel and holistic take on online self-presentation.

We begin in chapter 2 with a literature survey of fundamental ideas from social sciences about the presentation of self, as well as more recent digital sociology work from social and computer sciences, psychology, and media theory, about how people use such systems to express themselves and connect with others. Grounded in this, we discuss the current state of the art of Social Networking Sites (SNSs), and how they meet (or fail to meet) peoples' social needs.

In chapter 3 I present five self-contained empirical studies which allow us to analyse different aspects of online identity behaviours 'in the wild', and further discuss the concept of an online profile as a tool for self expression. I use the results of these studies to ground the description and justification of the conceptual framework.

Chapter 4 takes a brief look at the history of implementation and standardisation efforts for decentralised online social interactions. I use this review as a lead-in for a deeper look at the socio-technical process of formal standards development, in the context of a W3C Working Group in which I participated, in chapter 5. I also provide a prototype implementation of the standards produced by the group, as well as a design for a novel interaction pattern that can be used alongside.

In the concluding chapter I draw together these findings, and suggest directions for future research.

The contributions of this thesis have been generated by a combination of surveying existing theoretical and practical work; conducting empirical studies; and practice through writing and implementing Web standards.

Empirical studies take the form of descriptive [desc], whereby a detailed account of a particular situation is given; and ethnographic, whereby individual people are observed, surveyed, and interviewed [ethno]. These methods are a way of eliciting detailed insight into phenomena which create new ways of thinking about things or awareness of previously unknown possibilities, but do not necessarily provide a means to exhaustively map a problem space. For each individual study in chapter 3 I describe in more detail its particular method and limitations.

A conceptual framework is a useful cognitive tool, which "explains, either graphically or in narrative form, the main things to be studied - the key factors, concepts, or variables - and the presumed relationships among them." [miles94]. I developed mine through eliciting, aggregating and clustering many concepts from across my own study results and what I have learned from others.

Practice-led research is a way of immersing myself into the topic I'm studying, as well as directing outcomes and effecting change. Participation contributes a more complete understanding of an area than observation alone possibly can [ahrc07]. As such, I have been able to collaborate in designing standards, as well as report in detail on the process of standards-making. I have been able to build and use prototype systems, immersing myself in the perspectives of developer and user, and coming to a better understanding of the implications of both.