Recap of the World Wide Web 2013 Conference

As this was my second World Wide Web conference now (the first being Lyon, France, 2012), I has some expectations of what WWW13 would be like; in terms of the community, the papers, and the sheer energy of such a large conference. Rather than providing a day-by-day account of the conference and various activities, I’ve decided to highlight the aspects that stood out the most:

• The conference was organised brilliantly – lots of helpers, the registration process was well-thought out, and the use of the afternoons to enable those interested in visiting the local attractions (or sit on the beach) was a brilliant idea. Also, the use of the first two days, with tutorials in the morning and workshop filling the afternoon was great, although a full day of WWW (from 9am till 8pm) was quite long!
• The location of the conference made it very special indeed – Not only was it located in Rio, but the actual venue was brilliant. The Windsor Barra was large enough to accommodate all the WWW community, and offered plenty of large rooms for the running of the parallel sessions. I remember last year in Lyon, some of the rooms were overcrowded and not enough chairs were available for participants
• The variety and attendance of the workshops was great – Presenting at two of the workshops (SOCM and WOW), the community that attended them really provided a good atmosphere and environment for discussions. Rather than feeling like a mini-conference, the schedule, and formats of the workshops acted as a ‘talking-shop’ for new ideas, and developing those that had been presented. The two coffee breaks also supplemented this, providing those attending to go offline and further discuss (or debate, in some cases)! A large thanks needs to be given to the workshop organisers for putting in such a large amount of effort.
• The research tracks offered much variety – travelling with the Web Science DTC cohort from Southampton University (who included a mix set of students from different disciplinary backgrounds), there was a worry that the research might be too technical, however the variety of research topics and the skills of the presenters made the sessions within the research tracks accessible for all.
• The selection of keynote speakers and discussion panels really was a great addition to the already exciting schedule of research panels, demo sessions and poster presentations. Each day the participants were provided with a great opening keynote, and spoke about a variety of different topics and issues relating to not only the development of the WWW, but also the development of technology and science in general
• Overall, WWW2013 was a great success – excellent research, excellent people, and excellent location. The food was fantastic, and so was the traditional Brazilian dinner (and dance). Great Success! We were also offered a sneak preview of what WWW2014 in Seoul will be like; by the looks of the intro video, I’m sure it will be just as amazing!

With all that said, I’m going to talk a little more about the ideas and research that came out of the two workshops that I attended and presented at, both the Social Machines (SOCM) and Web Observatory (WOW) workshop. As these are very close to my PhD research and general areas of interest, I feel like they offered a great opportunity to share current research paradigms, directions and potentially future work and collaboration with the community.

Social Machines Workshop 2013

The first workshop, held on the Monday was SOCM, which was focusing on the growing variety of research surrounding the theory and practise of social machines (a term that received as much positive debate and discussion as the papers presented). The workshop, which was supported by the SOCIAM funding at Southampton University, provided a great forum for the current cutting edge and novel research that was involved in understanding how social machines of the Web (for now, let’s says a socio-technical web-based system) operate, and how we could potentially build and guide them. What was really interesting about the papers that were being presented (included my own) was the diversity of the theories and disciplines being drawn upon; to this point, Professor David De Roure made a great observation that this is potentially an evolution of Computer Science, from just designing systems in terms of the engineering, we are now dealing with a world where the design of the system is inherently human as it is technical, and as a result of this, we need methods, tools, and theory to understand how to better design these systems. However, unlike traditional design paradigms and approaches, we are no longer dealing with developing software or systems, at the micro level, the Web has offered us a platform to create something that is global, and as a result of this has many societal, economic and political implications.

This really brings me back to my research and interests, and the papers that were presented offered a great set of insights and discussion points in terms of what the Web actually is, and how can we understand it in terms of it being socio-technical, and what insights does that provide us (developers, individuals, or the collective ‘society’). There is also the issue of how does one define a social machine; what makes it social, and what makes it a ‘machine’? At the broadest of scales, everything could be considered a social machine, the interaction between a human and a hammer is inherently social and technical, not only in its use, but the societal needs for the hammer to be constructed, the shaping of the hammer because of the tools available, and the use of it (does a hammer have to be used just for nails…?). It is very easy to go down this route of saying everything is a social machine, so I think in the next few months we will be thinking very hard about what the boundaries and scope of this is. For now, I offer a few characteristics (which rose from the discussions surrounding my presentations) for what a social machine could be characterised as:

• Some Web-based characteristic or interaction, this might be native on the Web, i.e. a Web site, or an application, or it might be a community of practice that function and operate as a result of the Web (something like the Open Data community). I think this argument relates to the issue and constraints of scale
• The interaction and diffusion of information and (digital) data seems to be an important characteristic. In the human and hammer example, their interactions are socio-technical, but I don’t think it could be classed as a social machine. Perhaps a characteristic of a social machine needs be constrained to something that interacts (between humans or technologies), with digital data.
• A final point is the components and interactions of a social machine; can it be classified as a social machine if it is just technologies operating and interacting with each other (like in the example that David De Roure provided). There are two perspectives that could be taken,
  • The preposition that everything is socio-technical, i.e. the development of a technology has a human component (even if it is not visible, it was responsible for the design, and the social needs were embedded in its design). Then yes, interactions between two Web-technologies could be a social machine
  • The preposition that it is only a social machine if there is an active human component in its functionality and operation, i.e. mechanical Turk
  • I think the previous point is something to do with the adoption of the concept of socio-technical in terms of its application and understanding. This is something I’ve been trying to tackle for quite some time, how does one apply the concept of socio-technical to the Web, what does it actually provide in terms of description and analysis, and can it offer anything for those that are developing the Web (or Web activities, as I like to refer to them)? Thinking a little harder in terms of what socio-technical actually means is going to be a good thing for this type of research, and also for the broader WWW community as well.

These points are very much at the heart of the discussions of understanding what we mean by social machines, and through some more discussions, debates and empirically-driven research I think we will be much closer with a definition, a set of characteristics and potentially a framework to fully classify a spectrum of socio-technical systems that we consider as “social machines”.

Web Observatory Workshop 2012

The second workshop that I attended, WOW2013, focused on the efforts of those interested in the Web Observatory – an international WSTNet project which is currently gaining the support of those interested in providing a platform for academics, industry, researchers and government to store, and shares their data. This ties in very well with the research of social machines, as an essential part of being able to understand them and monitor them is to have a platform that can collect the variety and volume of the data that they generate. As like the previous workshop, the papers presented were of great variety, from the techniques and current approaches to creating a platform, the current usages of individual Web Observatories, and the incentives and design considerations of building and using them in a distributed manner.

The one aspect of this workshop which ties into my own research (and the section of the workshop that I was presenting in) was the use of different social machines to track and understand the diffusion of information and viral content on the Web. My research (which was the result of joint collaborative efforts with KAIST, South Korea) looked at how Wikipedia could be used as an indicator of soon-to-trend human activity, with a specific focus on how information is spread across different countries, and the implications of culture on this. The discussion which followed my presentation led to the potential collaboration with L3S, one of the partner WSTNet Labs in Germany. These kinds of outcomes and future collaborative work are only made possible by having these kinds of workshops along with the right community attending.

There was also an interesting discussion to be had with regards to Paul Booth and Paul Gaskell’s paper on looking at the exchange of value for data. At the most simplest level, they proposed the question of why would people want to share their data, what are their incentives and their rewards for doing it. A nice way to look at this is a market place for data; individuals who have datasets partake in this market place, the data – valued by some index and set of metrics – can be traded between individuals, which then creates a data index (like a stock market for data). This raises some questions of commercial vs. non-commercial markets, competition, or even more fundamental than this, how does one value data (especially as this will encounter a cold start problem). The concept of a data index and trading place is a great idea, however, I think we are going to have to think hard about how can a piece of data be valued, especially when it is first introduced into the network. Valuing it solely on the size of it, or the number of records has many issues, therefore the data on the market needs to have some specific meta-data associated with it in order to assign value to it. There is also the issue (and differentiation needed) between uploading a piece of data, downloading it, and actually using it; the latter being the most hard to quantity and track. Take data.gov.uk for instance, government departments can upload their data, then individuals can download it. However being able to track where this information is being used, and for what purpose (i.e. Applications, Website, etc.) is not simple. Their needs to be some technical intervention to enable some form of feedback loop for tracking data use.

There was also a lot of cross over and synergy with the discussions had at the previous day’s workshop on social machines and how the development of a Web Observatory will be an essential addition to understanding how social machines develop, how they evolve over time, etc. What is interesting to me is the ability to not only track these machines in terms of their macro characteristics (i.e. the network structures and the amount of data that they are providing), but also providing some way of understanding them at the micro level as well, in some sense, providing the meta-data around the captured quantitative data in order to add context to what is being observed. This goes back to the discussions regarding understanding these social machines in terms of the human activity, as well as the technological development. In order to understand their evolution (or even just the way they are being used), then we need a way to capture both the macro and micro, simultaneously. There is a lot he issue of how does one capture multiple social machines, and their interactions between them? What are the boundaries that exist between one social machine and other, and when capturing this data, how does one represent it? These are all questions that need a lot of further discussion, and I think it they will be answered not only by theorising, but by practise as well.

My training in the social sciences has told me that not everything is about building things, and I believe this is completely true. However, when we want to build something (which at some point, will be inevitable), understanding of the implications and interactions with the social is just as important as the technologies that underpin it. The construction of a social machine (if that is even possible), will social understanding, technological expertise, and tools such as the Web observatory to track and monitor its progress.

Final thoughts

The past week has been a great experience, catching up with those that I met last year in Lyon, and meeting a few new! It has also been an excellent time to discuss the ideas which we have all been interested in during the last 12 months here at Southampton. Not only has WWW2013 provided a forum for the cutting edge Web research, it has also been a great forum for Web Science discussion, which hopefully will continue to grow, especially with the WWW2014 Web Science track.