Telling stories to our computers
In the first two parts of this series looking at storytelling and technology, BBC User Experience Architect Paul Rissen examined the relationship between the author and the audience, and the ways in which stories were breaking free from traditional media. In his final essay, he speculates on the storytelling potential of data.
Discussions about storytelling and data tend to come in two forms at the moment. The first is the ‘big data‘ trend. Essentially, this boils down to the fact that immense processing power is relatively easy these days. As a result, we can chuck a shedload of data at a computer, apply statistical formulae, and try to detect patterns. Some call this ‘data storytelling’ – the idea that we can identify patterns in large amounts of data, and therefore provide insight into aspects of our lives. Of course, the problem with statistics and numbers is that they are inherently impersonal and abstract. They can also be twisted to fit the narrative which an author is trying to tell. The persuasive power of ‘infographics’ – numbers combined with arresting visuals, is both to be admired and feared.
The second aspect to the discussion around data is when it becomes ‘metadata’, nicely summed up with very good advice by Michael Bhaskar over on The Writing Platform. However, this too shapes our perceptions around what data is. Whereas in the ‘big data’ trend it’s all about numbers and the crunching thereof, here it’s about labelling and classification. Data is seen as a way of recreating and re-imagining the library experience – your creative work is, in effect, a black box, upon which you can attach descriptive labels: author, title, genre and so on. This leads to the perception that data is dry, boring, and certainly not creative.
Which is where I beg to differ. When someone sits down to watch, listen to or read a story, I’d be very surprised if they regarded the experience as being akin to handling a black box. Instead, the experience is all about opening up that box. It’s about entering the world, discovering new things, connecting the dots between the story in progress and the stories that have gone before. The black box, it transpires, is full of spiders, spinning their interconnected fibres. For me, this is the beauty and excitement of experiencing a story. And it seems crazy to me that when it comes to the Web, we tend to stop at the level of the black box.
Marshall McLuhan, who also made a brief appearance in the last article, talked about getting to grips with a new medium, not through the content, but the particular aspects and effects of its form. The medium is the message. When we look at the Web (not the Internet), in this light, what do we find?
Sir Tim Berners-Lee (creator of the Web, of course) was way ahead of us. He didn’t just envisage the Web as being a way of connecting documents together, but more importantly, it was about connecting the things inside those documents – the things that people really care about. To some extent, yes, this is about sharing identifiers for common concepts, but that doesn’t mean having to agree on a single truth. More importantly, the idea that Web identifiers (URIs) can be used to identify concepts, not just documents, and that the links that we make between these concepts can be more descriptive, semantic, is crucial. If we’re really serious about taking storytelling to the new medium of the Web, then it seems to me that we need to engage with what the form of the Web medium is: URIs to identify concepts, descriptive hyperlinks to represent the connections between those concepts.
We can, of course, combine this with the networked form and effects of the Internet, but for me, the Internet medium is about connecting computers and people, and passing messages between them in the form of bits and packets. When we consider the Web as the medium, it becomes about codifying meaning and perceived connections directly into the Web itself. Thus, data is no longer simply a mass of numbers or dry bibliographic information, but has the opportunity to be so much more – to directly express the creativity and beauty of our stories as data.
But why bother? Well, computers are stupid. There, I said it. No matter how smart we think they are, the truth is that in pretty much every computer system, the intelligence is in the logic, rules and models that one or many people have encoded into that system. Others have quite rightly pointed out, therefore, that we need to be careful about what ethics and philosophies we encode into our computer systems. But all you have to remember for now is that computers are stupid.
The world a computer sees is flat and meaningless. Whilst a fair number of us are employed as user experience designers, very few seem to care about the experience that the poor old computer is having, whilst ‘users’ are off being ‘delighted’. This might sound like folly, but a great deal of our frustrations with computers stem from the fact that the machines just don’t get it. Why should they? We can read the symbols ‘turkey’ and work out, from the context, whether we mean the bird or the country; to the computer, this remains a completely abstract linear series of symbols that it has learned to accept. Until now.
The idea of the robot-readable world has been bubbling away for a few years now. In the previous article in this series, I touched on one aspect of it, the world of connected devices. The other main thrust of the idea is that it is about learning to appreciate and build around how computers ‘see’ the world. If we can teach them to recognise certain things, and assign values to those things, then we begin to teach computers understanding. And once we’ve started doing that, designing things for humans, aided by computers, becomes a different proposition.
This is where the idea of the Web being used to express concepts comes in. It’s already beginning to take hold in the scientific and academic communities (though again, this tends to be the reserve of pretty dry, worthy data) – and it could be argued that developments such as Apple’s Siri service and Google’s Knowledge Graph rely, at least in part, on structured data being the raw material of the Web.
Perhaps one future of storytelling is one where anyone can interrogate and ask questions of the world of the story – where we don’t just have APIs for products and services, but every story is its’ own API, too (Jeni Tennison discusses a similar theme in terms of Government Data here). If a story can be released as text, audio, images and video, why not as data? Why not have an API that goes alongside the release of a new piece of fiction, regardless of medium?
It is possible to imagine a future where authors can lead audiences on certain paths across the story graph, but the universe of connected stories is open for people to explore, too. Stories could be more expansive than the concepts we hold in our minds, as, for instance, screen-based, encyclopaedic reference works power experiences connected across devices.
Perhaps if we start telling stories to our computers, instead of just expecting them to reveal stories to us from their flat, meaningless world, then we will finally find out what storytelling in the new medium of the networked, semantic Web might mean.