Building Webs of Discovery
Context: The folk at Unthinkable Digital asked me to write something about the dream & creative potential of linked open data and the Web, as I saw it, and what the barriers to it have been. The following is what resulted.
Although I don’t go into the practicalities of it in the piece, I do believe that through a well-thought-out, disciplined approach to information architecture & web design — basic things like one-URL-per-concept, and contextual links, based on domain models, between those concepts — such a dream is not beyond our grasp — or at the very least, worth experimenting with. It is, in effect, my current, long-term vision for the Web.
The hope of this piece, therefore, is to inspire people to imagine ways in which we can go beyond what we do with the Web currently. The more prosaic ‘how we actually do this’, and the limits of the approach, is something I’d happily discuss another time — but I would certainly advocate it as a viable approach to making things on the Web.
A dream of constellations
Clear your mind. Imagine a series of dots of light, in the black, dark emptiness of space. Now, imagine that each one of those dots was named for things in your life. You. Your family. Your friends. The objects you own. Your favourite TV shows. Your last holiday destination. The restaurant you visited with friends to celebrate your birthday. Each significant moment in your life.
Now, imagine streams of light, connecting each of those dots, connecting them in meaningful ways that mirror the connections between you and your friends; the ways in which TV shows and films share the same actors; or the cause and effect relationships you might choose to draw between moments in your life. These connections carry within themselves not just a simple back-and-forth message of what each connected star represents, but a description of the connection itself — sisters, cause, effect, authorship, and so on.
These constellations of light are networks that, unlike the constellations in our sky, do not necessarily form meaningful patterns when seen at a distance, but whose meaning is in the stars and connections themselves.
These constellations currently exist only in your mind — if you’re lucky, and particularly good at communication, they may exist in others’ minds, too. At its heart, any form of human communication is about trying to reproduce copies of these constellations of concepts, transmitting them from one human mind to another.
This is what the World Wide Web could be. A way for us to create constellations of concepts, patterns of connections between things that allow us to share ideas, to hold them in a form that multiple people can point at, look at, examine, interrogate and riff off, without the lost-in-translation misunderstandings that mire so many attempts at audiovisual communication between us.
This, at a high and possibly overly-grandiose level, is the aim of what Sir Tim Berners-Lee calls the ‘Giant Global Graph’, the next stage on from the World Wide Web. The Internet was about connecting machines together; the World Wide Web connects documents; the Giant Global Graph would connect ideas (or, as Berners-Lee puts it, ‘the things that those documents are about’).
Beyond the comfort zone
We don’t tend to think of this, at all, in our daily lives. We’re quite happy conflating the Web and the Internet, and talking about ‘pages’ and ‘streams’ — the computer as simulation of paper documents, and other existing media forms; the Internet as a distribution mechanism, giving us an endless flow of this ‘content’. For as much as we have hyperlinks between these documents, each of those things is a silo, containing multitudes of concepts, connections, and ideas.
At the same time, if you look just a little deeper, we do instinctively get the idea of connecting concepts — think of tagging your friends in photos on Facebook, or using hashtags to link otherwise disconnected tweets about the same thing. The hashtag is an agreement, a phrase which people can use to mean the same idea. Hashtags link tweets, but nothing, yet, links the hashtags themselves. The nearest we have to a network of concepts, is again, something we’re familiar with, even if we don’t yet grasp its potential — social networks — people, represented on the Web, and linked together by their real world relationships. This is the reality of today’s ‘Giant Global Graph’, or, as it’s sometimes also known, the Semantic Web.
It’s a source of deep frustration to many, including myself, that we seem to have suffered a collective failure of imagination. Ever since Twitter, Facebook, Instagram and Tumblr came on the scene, we seem to have stopped talking about the greater potential of these webs of meaning, and lost ourselves in becoming ‘content creators’, endlessly speculating on the potential and effects of these existing social media, without seeing that there is so much more out there to create.
Why should we be restricted to only thinking about the links between pages, between documents and photos, between Youtube videos and Soundcloud tracks? Why shouldn’t we see the building of constellations that represent our social connections, in Facebook, Twitter and the like, or between songs in iTunes, Spotify, Google Play Music, as only the beginning — that the Web doesn’t have to just be about sharing and connecting existing media forms, but about creating, connecting and sharing webs, constellations, networks, graphs, call them what you will, as creative works in and of themselves?
I’d like to see a world where every creative work comes with its own graph, which in turn can be connected to other graphs across the Web. Or, even better, just as we create a painting, a novel, a video or piece of music, creating a graph should be seen as work of art — a work of art with the inherent properties of the Web — it can be traversed in many ways, not only through a single, creator-defined path; the ideas and concepts within can be connected to other creative work graphs that share the same, or similar, elements — think of the endless playful connections between films, TV shows and plays all set in the same era; finally, each node in a creative work graph would be a point not only of connection, but a seed from which other, new works could be created — networked music mashups, a constellation of connected fan fiction, perhaps.
Imagination vs iPlayer
So why aren’t we there yet? There are, unfortunately, several reasons why, in my opinion.
Firstly, yes, the failure of imagination described above. It’s understandable — the Web is not even thirty years old, the use of the Internet in mainstream popular culture, not much older. We’re at a point now where it seems that the lustre of the social networks is finally wearing off, but what comes next is still unknown.
For now, the world remains focused on the sheer novelty of a medium that can represent existing media forms in a way that can be perfectly copied and shared almost instantaneously, for better and for worse.
It wasn’t always this way — before the launch of iPlayer, a small team at the BBC saw the potential of what could be done with the Web, and this lead to many great projects — the launch of the /programmes and /music platforms, but the success of iPlayer, as excellent as it was, blinded the industry. It appears that just distributing existing forms of media, or looking to personalisation and virtual reality, would be enough to count as innovation. Perhaps it is — but I must admit feeling sad that this alternative, rich seam of treating the Web as a medium itself, was seemingly abandoned so quickly.
Similarly, we suffer from a lack of vision for the potential of computers — recreating, and quite rightly improving wherever possible, existing services and processes from our analogue world. Innovation comes about through dreaming up new ways to exchange money, or to use the sheer brute force of computing power to run algorithms over our behaviour, to tell us what we already know, or couldn’t quite see yet. But all the big data and neural networks in the world won’t help us grasp that we have a medium for creative expression staring us right in the face.
The map is not the territory
Secondly, those who did see the potential of the Semantic Web, in its earlier days have themselves been blinded by the sheer complexity that lies under the surface of the picture I painted at the start of this piece. There are seriously fiendish theoretical, logical and computational questions to be asked about the representation and storage of such a conceptual model of human thought — but these have often put a damper on any practical implementation, however lightweight, of the poetic potential of these creative constellations.
Storage and complex philosophical debates aside, just publishing the basics as structured, linked, and ideally open (either openly licensed, openly available, or both), data, would be a huge win. For the Semantic Web, storage of these graphs was never the point — if a relational database or document store is good for you, use it — but publishing the data, sharing your creation with the world — that’s where the value lies.
All too often, this way of seeing what we could do with the Web has been conflated with these theoretical, academic and deeply obtuse questions, so much so that the simple, creative potential has been forgotten — the whole approached dismissed as boring, technical, stuffy, and overly complex, not something for creative minds to engage with. Best for them to stick to ‘real’ design, design that exists on screens, not in the realm of imagination. Ted Nelson’s view of computers as ‘literary machines’, meanwhile, waits in the wings.
Linked Open Data, has, for the most part, been an academic exercise, releasing huge datasets, yes, but mainly in the realms of science and administration. BBC Music, Programmes and even the TV Tropes database are shining, yet spluttering lights in the graph, as their potential goes unremarked, and their existence threatened by changing business models and well-intentioned, yet unwieldy, API management practices.
Glimmers in the night sky
Yet, there is light in the dark. Google have been quietly building up their Knowledge Graph, and successfully demonstrating practical, user-oriented applications of networks of structured data. DBPedia, a structured data version of Wikipedia, hovers on the border between academia and empowering real user experiences, such as the BBC’s Shakespeare Archive Resource for education. Facebook Open Graph, Schema.org, even voice interfaces such as Amazon’s Alexa, Apple’s Siri and Google Assistant, are getting us used to interacting directly with structured data and information, without the distraction of pages or screens.
But as long as we rely on private corporations to lead us into this space, we run the very real risk of signing away our lives and rights, not just our documents, videos and photos. It’s up to us to understand the Web, to own our own data, and to educate ourselves about the complexities that come from putting everything we know about our worlds into a shared, interlinked arena.
Practical applications
More importantly, we’re beginning to see the limitations of a document-based and ‘dumb’ hyperlinked world. Discussions are now turning to the need to establish trust, authenticity, provenance — to find understanding within this morass of content and ill-defined hyperlinks. Social interventions, attempts to change our behaviour, will play an important role if we are to survive in turbulent times, and the machines are waiting here, to help us, if only we grasp their potential.
The machines are, of course, not a silver bullet. Pushing semantics into the Web can just as easily be used to distract as well as inform, and there will be difficulties, such as the danger of overreach: representing and connecting our understanding of each other and our world is complex enough, without trying to replicate actual reality. Not every problem in the world can, or should, be solved through technology — sometimes it takes communities to work together to enact real social, cultural and political change.
But the focus on the difficulties should not deter us from making what we can of the medium at our disposal — small, useful, descriptive networks of ideas, ones that help us recognise the need for what Robert Anton Wilson called multi-model agnosticism — being humble about what you know, being open to other ways of understanding the world, and accepting that multiple conceptual models can exist concurrently. Webs that help us recognise that multiple points of view exist — that not everything is cut and dried; that under-represented communities have the right to express their ideas, too — whilst refocusing and doubling down on the need for provenance, for evidence, to be always asking questions.
Can linked data save journalism?
To see that perhaps, some future version of Google’s PageRank weighting system can help deter flimsy, fact-less misinformation, discrediting attempts to warp our understanding of the world, and boosting well researched, cited and linked news stories instead. Perhaps the future of news is in reputation building through trust, showing your workings, allowing the audience to interrogate and explore the news, such as with Membrane, a New York Times experiment by Jane Friedhoff — to build a well-informed, media literate public that holds power to account.
…and bring us creative renewal?
And perhaps, somewhere in that future, creativity is no longer restricted to creating audio, video, text, or interactive combinations of those traditional media — but to creating structured data, given addressable, point-at-able, connectable, mashup-able, constellations of ideas. Constellations that can find their instantiation in multiple media, be accessible to all, and lead to better communication between all of us.
The murky depths of logic and machines do lead to complex issues. But that shouldn’t stop us looking up at the stars, and creating our own networks of light. It’s not difficult, either. All it needs is for people to realise that the Web doesn’t have to just be pages. It can be, and enable, so much more.
Thanks to Matthew Shorter, at Unthinkable Digital, for commissioning and copy-editing this.