The Soul of Web 2.0

Introduction: In The Beginning

Back in the 1980s, when personal computers mostly meant IBM PCs running Lotus 1*2*3 and, perhaps, if you were a bit off-center, an Apple Macintosh running Aldus Pagemaker, the idea of a coherent and interconnected set of documents spanning the known human universe seemed fanciful.  But there have always been dreamers, among them such luminaries as Douglas Engelbart, who gave us the computer mouse, and Ted Nelson, who coined the word ‘hypertext’.  Engelbart demonstrated a fully-functional hypertext system in December 1968, the famous ‘Mother of all Demos’, which framed computing for the rest of the 20th century.  Before man had walked on the Moon, before there was an Internet, we had a prototype for the World Wide Web.  Nelson took this idea and ran with it, envisaging a globally interconnected hypertext system, which he named ‘Xanadu’ – after the poem by Coleridge – and which attracted a crowd of enthusiasts intent on making it real.  I was one of them.  From my garret in Providence, Rhode Island, I wrote a front end – a ‘browser’ if you will – to the soon-to-be-released Xanadu.  This was back in 1986, nearly five years before Tim Berners-Lee wrote a short paper outlining a universal protocol for hypermedia, the basis for the World Wide Web.

Xanadu was never released, but we got the Web.  It wasn’t as functional as Xanadu – copyright management was a solved problem with Xanadu, whereas on the Web it continues to bedevil us – and links were two-way affairs; you could follow the destination of a link back to its source.  But the Web was out there and working for thousand of people by the middle of 1993, while Xanadu, shuffled from benefactor to benefactor, faded and finally died.  The Web was good enough to get out there, to play with, to begin improving, while Xanadu – which had been in beta since the late 1980s – was never quite good enough to be released.  ‘The Perfect is the Enemy of the Good’, and nowhere is it clearer than in the sad story of Xanadu.

If Xanadu had been released in 1987, it would have been next to useless without an Internet to support it, and the Internet was still very tiny in the 1980s.  When I started using the Internet, in 1988, the main trunk line across the United States was just about to be upgraded from 9.6 kilobits to 56 kilobits.  That’s the line for all of the traffic heading from one coast to the other.  I suspect that today this cross-country bandwidth, in aggregate, would be measured in terabits – trillions of bits per second, a million-fold increase.  And it keeps on growing, without any end in sight.

Because of my experience with Xanadu, when I first played with NCSA Mosaic – the first publicly available Web browser – I immediately knew what I held in my mousing hand.  And I wasn’t impressed.  In July 1993 very little content existed for the Web – just a handful of sites, mostly academic.  Given that the Web was born to serve the global high-energy-physics community headquartered at CERN and Fermilab, this made sense.  I walked away from the computer that July afternoon wanting more.  Hypertext systems I’d seen before.  What I lusted after was a global system with a reach like Xanadu.

Three months later, when I’d acquired a SUN workstation for a programming project, I immediately downloaded and installed NCSA Mosaic, to find that the Web elves had been busy.  Instead of a handful of sites, there were now hundreds.  There was a master list of known sites, maintained at NCSA, and over the course of a week in October, I methodically visited every site in the list.  By Friday evening I was finished.  I had surfed the entire Web.  It was even possible to keep up the new sites as they were added to the bottom of the list, though the end of 1993.  Then things began to explode.

From October on I became a Web evangelist.  My conversion was complete, and my joy in life was to share my own experience with my friends, using my own technical skills to get them set up with Internet access and their own copies of NCSA Mosaic.  That made converts of them; they then began to work on their friends, and so by degrees of association, the word of the Web spread.

In mid-January 1994, I dragged that rather unwieldy SUN workstation across town to show it off at a house party / performance event known as ‘Anon Salon’, which featured an interesting cross-section of San Francisco’s arts and technology communities.  As someone familiar walked in the door at the Salon, I walked up to them and took them over to my computer.  “What’s something you’re interested in?” I’d ask.  They’d reply with something like “Gardening” or “Astronomy” or “Watersports of Mesoamerica” and I’d go to the newly-created category index of the Web, known as Yahoo!, and still running out of a small lab on the Stanford University campus, type in their interest, and up would come at least a few hits.  I’d click on one, watch the page load, and let them read.  “Wow!” they’d say.  “This is great!”

I never mentioned the Web or hypertext or the Internet as I gave these little demos.  All I did was hook people by their own interests.  This, in January 1994 in San Francisco, is what would happen throughout the world in January 1995 and January 1996, and still happening today, as the two-billion Internet-connected individuals sit down before their computers and ask themselves, “What am I passionate about?”

This is the essential starting point for any discussion of what the Web is, what it is becoming, and how it should be presented.  The individual, with their needs, their passions, their opinions, their desires and their goals is always paramount.  We tend to forget this, or overlook it, or just plain ignore it.  We design from a point of view which is about what we have to say, what we want to present, what we expect to communicate.  It’s not that that we should ignore these considerations, but they are always secondary.  The Web is a ground for being.  Individuals do not present themselves as receptacles to be filled.  They are souls looking to be fulfilled.  This is as true for children as for adults – perhaps more so – and for this reason the educational Web has to be about space and place for being, not merely the presentation of a good-looking set of data.

How we get there, how we create the space for being, is what we have collectively learned in the first seventeen years of the web.  I’ll now break these down some of these individually.

I: Sharing

Every morning when I sit down to work at my computer, I’m greeted with a flurry of correspondence and communication.  I often start off with the emails that have come in overnight from America and Europe, the various mailing lists which spit out their contents at 3 AM, late night missives from insomniac friends, that sort of thing.  As I move through them, I sort them: this one needs attention and a reply, this one can get trashed, and this one – for one reason or another – should be shared.  The sharing instinct is innate and immediate.  We know upon we hearing a joke, or seeing an image, or reading an article, when someone else will be interested in it.  We’ve always known this; it’s part of being a human, and for as long as we’ve been able to talk – both as children and as a species – we’ve babbled and shared with one another.  It’s a basic quality of humanity.

Who we share with is driven by the people we know, the hundred-and-fifty or so souls who make up our ‘Dunbar Number’, the close crowd of individuals we connect to by blood or by friendship, or as co-workers, or neighbors, or co-religionists, or fellow enthusiasts in pursuit of sport or hobby.  Everyone carries that hundred and fifty around inside of them.  Most of the time we’re unaware of it, until that moment when we spy something, and immediately know who we want to share it with.  It’s automatic, requires no thought.  We just do it.

Once things began to move online, and we could use the ‘Forward’ button on our email clients, we started to see an acceleration and broadening of this sharing.  Everyone has a friend or two who forwards along every bad joke they come across, or every cute photo of a kitten.  We’ve all grown used to this, very tolerant of the high level of randomness and noise, because the flip side of that is a new and incredibly rapid distribution medium for the things which matter to us.  It’s been truly said that ‘If news is important, it will find me,’ because once some bit of information enters our densely hyperconnected networks, it gets passed hither-and-yon until it arrives in front of the people who most care about it.

That’s easy enough to do with emails, but how does that work with creations that may be Web-based, or similarly constrained?  We’ve seen the ‘share’ button show up on a lot of websites, but that’s not the entire matter.  You have to do more than request sharing.  You have to think through the entire goal of sharing, from the user’s perspective.  Are they sharing this because it’s interesting?  Are they sharing this because they want company?  Are they sharing this because it’s a competition or a contest or collaborative?  Or are they only sharing this because you’ve asked them to?

Here we come back – as we will, several more times – to the basic position of the user’s experience as central to the design of any Web project.  What is it about the design of your work that excites them to share it with others?  Have you made sharing a necessary component – as it might be in a multi-player game, or a collaborative and crowdsourced knowledge project – or is it something that is nice but not essential?  In other words, is there space only for one, or is there room to spread the word?  Why would anyone want to share your work?  You need to be able to answer this: definitively, immediately, and conclusively, because the answer to that question leads to the next question.  How will your work be shared?

Your works do not exist in isolation.  They are part of a continuum of other works?  Where does your work fit into that continuum?  How do the instructor and student approach that work?  Is it a top-down mandate?  Or is it something that filters up from below as word-of-mouth spreads?  How does that word-of-mouth spread?

Now you have to step back and think about the users of your work, and how they’re connected.  Is it simply via email – do all the students have email addresses?  Do they know the email addresses of their friends?  Or do you want your work shared via SMS?  A QRCode, perhaps?  Or Facebook or Twitter or, well, who knows?  And how do you get a class of year 3 students, who probably don’t have access to any of these tools, sharing your work?

You do want them to share, right?

This idea of sharing is foundational to everything we do on the Web today.  It becomes painfully obvious when it’s been overlooked.  For example, the iPad version of The Australian had all of the articles of the print version, but you couldn’t share an article with a friend.  There was simply no way to do that.  (I don’t know if this has changed recently.)  That made the iPad version of The Australian significantly less functional than its website version – because there I could at least past a URL into an email.

The more something is shared, the more valuable it becomes.  The more students use your work, the more indispensable you become to the curriculum, and the more likely your services will be needed, year after year, to improve and extend your present efforts.  Sharing isn’t just good design, it’s good business.

II: Connecting

Within the space for being created by the Web, there is room for a crowd.  Sometimes these crowds can be vast and anonymous – Wikipedia is a fine example of this.  Everyone’s there, but no one is wholly aware of anyone else’s presence.  You might see an edit to a page, or a new post on the discussion for a particular topic, but that’s as close as people come to one another.  Most of the connecting for the Wikipedians – the folks who behind-the-scenes make Wikipedia work – is performed by that old reliable friend, email.

There are other websites which make connecting the explicit central point of their purpose.  These are the social networks: Facebook, MySpace, LinkedIn, and so on.  In essence they take the Dunbar Number written into each of our minds and make it explicit, digital and a medium for communication.  But it doesn’t end there; one can add countless other contacts from all corners of life, until the ‘social graph’ – that set of connections – becomes so broad it is essentially meaningless.  Every additional contact makes the others less meaningful, if only because there’s only so much of you to go around.

That’s one type of connecting.  There is another type, as typified by Twitter, in which connections are weaker – generally falling outside the Dunbar Number – but have a curious resilience that presents unexpected strengths.  Where you can poll your friends on Facebook, on Twitter you can poll a planet.  How do I solve this problem?  Where should I eat dinner tonight?  What’s going on over there?  These loose but far-flung connections provide a kind of ‘hive mind’, which is less precise, and knows less about you, but knows a lot more about everything else.

These are not mutually exclusive principles.  It’s is not Facebook-versus-Twitter; it is not tight connections versus loose connections.  It’s a bit of both.  Where does your work benefit from a tight collective of connected individuals?  Is it some sort of group problem-solving?  A creative activity that really comes into its own when a whole band of people play together?  Or simply something which benefits from having a ‘lifeline’ to your comrades-in-arms?  When you constantly think of friends, that’s the sort of task that benefits from close connectivity.

On the other hand, when you’re collaborating on a big task – building up a model or a database or an encyclopedia or a catalog or playing a massive, rich, detailed and unpredictable game, or just trying to get a sense of what is going on ‘out there’, that’s the kind of task which benefits from loose connectivity.  Not every project will need both kinds of connecting, but almost every one will benefit from one or the other.  We are much smarter together than individually, much wiser, much more sensible, and less likely to be distracted, distraught or depressed.  (We are also more likely to reinforce each others’ prejudices and preconceptions, but that’s another matter of longstanding which technology can not help but amplify.)  Life is meaningful because we, together, give it meaning.  Life is bearable because we, together, bear the load for one another.  Human life is human connection.

The Web today is all about connecting.  That’s its single most important feature, the one which is serving as an organizing principle for nearly all activity on it.  So how do your projects allow your users to connect?  Does your work leave them alone, helpless, friendless, and lonely?  Does it crowd them together into too-close quarters, so that everyone feels a bit claustrophobic?  Or does it allow them to reach out and forge the bonds that will carry them through?

III: Contributing, Regulating, Iterating

In January of 2002, when I had my first demo of Wikipedia, the site had barely 14,000 articles – many copied from the 1911 out-of-copyright edition of Encyclopedia Britannica.  That’s enough content for a child’s encyclopedia, perhaps even for a primary school educator, but not really enough to be useful for adults, who might be interested in almost anything under the Sun.  It took the dedicated efforts of thousands of contributors for several years to get Wikipedia to the size of Britannica (250,000 articles), an effort which continues today.

Explicit to the design of Wikipedia is the idea that individuals should contribute.  There is an ‘edit’ button at the top of nearly every page, and making changes to Wikipedia is both quick and easy.  (This leaves the door open a certain amount of childish vandalism, but that is easily reversed or corrected precisely because it is so easy to edit anything within the site.)  By now everyone knows that Wikipedia is the collaboratively created encyclopedia, representing the best of all of what its contributors have to offer.  For the next hundred years academics and social scientists will debate the validity of crowdsourced knowledge creation, but what no one can deny is that Wikipedia has become an essential touchstone, our common cultural workbook.  This is less because of Wikipedia-as-a-resource than it is because we all share a sense of pride-in-ownership of Wikipedia.  Probably most of you have made some small change to Wikipedia; a few of you may have authored entire articles.  Every time any of us adds our own voice to Wikipedia, we become part of it, and it becomes part of us.  This is a powerful logic, an attraction which transcends the rational.  People cling to Wikipedia – right or wrong – because it is their own.

It’s difficult to imagine a time will come when Wikipedia will be complete.  If nothing else, events continue to occur, history is made, and all of this must be recorded somewhere in Wikipedia.  Yet Wikipedia, in its English-language edition, is growing more slowly in 2010 than in 2005.  With nearly 3.5 million articles in English, it’s reasonably comprehensive, at least by its own lights.  Certain material is considered inappropriate for Wikipedia – homespun scientific theories, or the biographies of less-than-remarkable individuals – and this has placed limits on its growth.  It’s possible that within a few years we will regard Wikipedia as essentially complete – which is, when you reflect upon it, an utterly awesome thought.  It will mean that we have captured the better part of human knowledge in a form accessible to all.  That we can all carry the learned experience of the species around in our pockets.

Wikipedia points to something else, quite as important and nearly as profound: the Web is not ‘complete’.  It is a work-in-progress.  Google understands this and releases interminable beta versions of every product.  More than this, it means that nothing needs to offer all the answers.  I would suggest that nothing should offer all the answers.  Leaving that space for the users to add what they know – or are willing to learn – to the overall mix creates a much more powerful relationship with the user, and – counterintuitively – with less work from you.  It is up to you to provide the framework for individuals to contribute within, but it is not up to you to populate that framework with every possibility.  There’s a ‘sweet spot’, somewhere between nothing and too much, which shows users the value of contributions but allows them enough space to make their own.

User contributions tend to become examples in their own right, showing other users how it’s done.  This creates a ‘virtuous cycle’ of contributions leading to contributions leading to still more contributions – which can produce the explosive creativity of a Wikipedia or TripAdvisor or an eBay or a RateMyProfessors.com.

In each of these websites it needs to be noted that there is a possibility for ‘bad data’ to work its way into system.   The biggest problem Wikipedia faces is not vandalism but the more pernicious types of contributions which look factual but are wholly made up.  TripAdvisor is facing a class-action lawsuit from hoteliers who have been damaged by anonymous negative ratings of their establishments.  RateMyProfessors.com is the holy terror of the academy in the United States.  Each of these websites has had to design systems which allow for users to self-regulate peer contributions.  In some cases – such as on a blog – it’s no more than a ‘report this post’ button, which flags it for later moderation.  Wikipedia promulgated a directive that strongly encouraged contributors to provide a footnote linking to supporting material.  TripAdvisor gives anonymous reviewers a lower ranking.  eBay forces both buyers and sellers to rate each transaction, building a database of interactions which can be used to guide others when they come to trade.  Each of these are social solutions to social problems.

Web2.0 is not a technology.  It is a suite of social techniques, and each technique must be combined with a social strategy for deployment, considering how the user will behave: neither wholly good nor entirely evil.  It is possible to design systems and interfaces which engage the better angels of nature, possible to develop wholly open systems which self-regulate and require little moderator intervention.  Yet it is not easy to do so, because it is not easy to know in advance how any social technique can be abused by those who employ it.

This means that aWeb2.0 concept that should guide you in your design work is iteration.  Nothing is ever complete, nor ever perfect.  The perfect is the enemy of the good, so if you wait for perfection, you will never release.  Instead, watch your users, see if they struggle to work within the place you have created for then, or whether they immediately grasp hold and begin to work.  In their more uncharitable moments, do they abuse the freedoms you have given them?  If so, how can you redesign your work, and ‘nudge’ them into better behavior?  It may be as simple as a different set of default behaviors, or as complex as a set of rules governing a social ecosystem.  And although Moses came down from Mount Sinai with all ten commandments, you can not and should not expect to get it right on a first pass.  Instead, release, observe, adapt, and re-release.  All releases are soft releases, everything is provisional, and nothing is quite perfect.  That’s as it should be.

IV: Opening

Two of the biggest Web2.0 services are Facebook and Twitter.  Although they seem to be similar, they couldn’t be more different.  Facebook is ‘greedy’, hoarding all of the data provided by its users, all of their photographs and conversations, keeping them entirely for itself.  If you want to have access to that data, you need to work with Facebook’s tools, and you need to build an application that works within Facebook – literally within the web page.  Facebook has control over everything you do, and can arbitrarily choose to limit what you do, even shut you down your application if they don’t like it, or perceive it as somehow competitive with Facebook.  Facebook is entirely in control, and Facebook holds onto all of the data your application needs to use.

Twitter has taken an entirely different approach.  From the very beginning, anyone could get access to the Twitter feed – whether for a single individual (if their stream of Tweets had been made public), or for all of Twitter’s users.  Anyone could do anything they wanted with these Tweets – though Twitter places restrictions on commercial re-use of their data.  Twitter provided very clear (and remarkably straightforward) instruction on how to access their data, and threw the gates open wide.

Although Facebook has half a billion users, Twitter is actually more broadly used, in more situations, because it has been incredibly easy for people to adapt Twitter to their tasks.  People have developed computer programs that send Tweets when the program is about to crash, created vast art projects which allow the public to participate from anywhere around the world, or even a little belt worn by a pregnant woman which sends out a Tweet every time the baby kicks!  It’s this flexibility which has made Twitter a sort of messaging ‘glue’ on the Internet of 2010, and that’s something Facebook just can’t do, because it’s too closed in upon itself.  Twitter has become a building block: when you write a program which needs to send a message, you use Twitter.  Facebook isn’t a building block.  It’s a monolith.

How do you build for openness?  Consider: another position the user might occupy is someone trying to use your work as a building block within their own project.  Have you created space for your work to be re-used, to be incorporated, to be pieced apart and put back together again?  Or is it opaque, seamless, and closed?  What about the data you collect, data the user has generated?  Where does that live?  Can it be exported and put to work in another application, or on another website?  Are you a brick or are you a brick wall?

When you think about your design – both technically and from the user’s experience – you must consider how open you want to be, and weigh the price of openness (extra work, unpredictability) against the price of being closed (less useful).  The highest praise you can receive for your work is when someone wants to use it in their own. For this to happen, you have to leave the door open for them.  If you publish the APIs to access the data you collect; if you build your work modularly, with clearly defined interfaces; if you use standards such as RSS and REST where appropriate, you will create something that others can re-use.

One of my favorite lines comes from science fiction author William Gibson, who wrote, ‘The street finds its own uses for things – uses the manufacturer never imagined.’  You can’t know how valuable your work will be to someone else, what they’ll see in it that you never could, and how they’ll use it to solve a problem.

All of these techniques – sharing, connecting, contributing, regulating, iterating and opening – share a common thread: they regard the user’s experience as paramount and design as something that serves the user.  These are not precisely the same Web2.0 domains others might identify.  That’s because Web2.0 has become a very ill-defined term.  It can mean whatever we want it to mean.  But it always comes back to experience, something that recognizes the importance and agency of the user, and makes that the center of the work.

It took us the better part of a decade to get to Web2.0; although pieces started showing up in the late 1990s, it wasn’t until the early 21st century that we really felt confident with the Web as an experience, and could use that experience to guide us into designs that left room for us to explore, to play and to learn from one another.  In this decade we need to bring everything we’ve learned to everything we create, to avoid the blind traps and dead ends of a design which ignores the vital reality of the people who work with what we create.  We need to make room for them.  If we don’t, they will make other rooms, where they can be themselves, where they can share what they’ve found, connect with the ones they care about, collaborate and contribute and create.

What Ever Happened to the Book?

For Ted Nelson

I: Centrifugal Force

We live in the age of networks.  Wherever we are, five billion of us are continuously and ubiquitously connected.  That’s everyone over the age of twelve who earns more than about two dollars a day.  The network has us all plugged into it.  Yet this is only the more recent, and more explicit network.  Networks are far older than this most modern incarnation; they are the foundation of how we think.  That’s true at the most concrete level: our nervous system is a vast neural network.  It’s also true at a more abstract level: our thinking is a network of connections and associations.  This is necessarily reflected in the way we write.

I became aware of this connectedness of our thoughts as I read Ted Nelson’s Literary Machines back in 1982.  Perhaps the seminal introduction to hypertext, Literary Machines opens with the basic assertion that all texts are hypertexts.  Like it or not, we implicitly reference other texts with every word we write.  It’s been like this since we learned to write – earlier, really, because we all crib from one another’s spoken thoughts.  It’s the secret to our success.  Nelson wanted to build a system that would make these implicit relationships explicit, exposing all the hidden references, making text-as-hypertext a self-evident truth.  He never got it.  But Nelson did influence a generation of hackersSir Tim Berners-Lee among them – and pushed them toward the implementation of hypertext.

As the universal hypertext system of HTTP and HTML conquered all, hypertext revealed qualities as a medium which had hitherto been unsuspected.  While the great strength of hypertext is its capability for non-linearity – you can depart from the text at any point – no one had reckoned on the force (really, a type of seduction) of those points of departure.  Each link presents an opportunity for exploration, and is, in a very palpable sense, similar to the ringing of a telephone.  Do we answer?  Do we click and follow?  A link is pregnant with meaning, and passing a link by necessarily incurs an opportunity cost.  The linear text is constantly weighed down with a secondary, ‘centrifugal’ force, trying to tear the reader away from the inertia of the text, and on into another space.  The more heavily linked a particular hypertext document is, the greater this pressure.

Consider two different documents that might be served up in a Web browser.  One of them is an article from the New York Times Magazine.  It is long – perhaps ten thousand words – and has, over all of its length, just a handful of links.  Many of these links point back to other New York Times articles.  This article stands alone.  It is a hyperdocument, but it has not embraced the capabilities of the medium.  It has not been seduced.  It is a spinster, of sorts, confident in its purity and haughty in its isolation.  This article is hardly alone.  Nearly all articles I could point to from any professional news source portray the same characteristics of separateness and resistance to connect with the medium they employ.  We all know why this is: there is a financial pressure to keep eyes within the website, because attention has been monetized.  Every link presents an escape route, and a potential loss of income.  Hence, links are kept to a minimum, the losses staunched.  Disappointingly, this has become a model for many other hyperdocuments, even where financial considerations do not conflict with the essential nature of the medium.  The tone has been set.

On the other hand, consider an average article in Wikipedia.  It could be short or long – though only a handful reach ten thousand words – but it will absolutely be sprinkled liberally with links.  Many of these links will point back into Wikipedia, allowing someone to learn the meaning of a term they’re unfamiliar with, or explore some tangential bit of knowledge, but there also will be plenty of links that face out, into the rest of the Web.  This is a hyperdocument which has embraced the nature of medium, which is not afraid of luring readers away under the pressure of linkage.  Wikipedia is a non-profit organization which does not accept advertising and does not monetize attention.  Without this competition of intentions, Wikipedia is itself an example of another variety of purity, the pure expression of the tension between the momentum of the text and centrifugal force of hypertext.

Although commercial hyperdocuments try to fence themselves off from the rest of the Web and the lure of its links, they are never totally immune from its persistent tug.  Just because you have landed somewhere that has a paucity of links doesn’t constrain your ability to move non-linearly.  If nothing else, the browser’s ‘Back’ button continually offers that opportunity, as do all of your bookmarks, the links that lately arrived in email from friends or family or colleagues, even an advertisement proffered by the site.  In its drive to monetize attention, the commercial site must contend with the centrifugal force of its own ads.  In order to be situated within a hypertext environment, a hyperdocument must accept the reality of centrifugal force, even as it tries, ever more cleverly, to resist it.  This is the fundamental tension of all hypertext, but here heightened and amplified because it is resisted and forbidden.  It is a source of rising tension, as the Web-beyond-the-borders becomes ever more comprehensive, meaningful and alluring, while the hyperdocument multiplies its attempts to ensnare, seduce, and retain.

This rising tension has had a consequential impact on the hyperdocument, and, more broadly, on an entire class of documents.  It is most obvious in the way we now absorb news.  Fifteen years ago, we spread out the newspaper for a leisurely read, moving from article to article, generally following the flow of the sections of the newspaper.  Today, we click in, read a bit, go back, click in again, read some more, go back, go somewhere else, click in, read a bit, open an email, click in, read a bit, click forward, and so on.  We allow ourselves to be picked up and carried along by the centrifugal force of the links; with no particular plan in mind – except perhaps to leave ourselves better informed – we flow with the current, floating down a channel which is shaped by the links we encounter along the way.  The newspaper is no longer a coherent experience; it is an assemblage of discrete articles, each of which has no relation to the greater whole.  Our behavior reflects this: most of us already gather our news from a selection of sources (NY Times, BBC, Sydney Morning Herald and Guardian UK in my case), or even from an aggregator such as Google News, which completely abstracts the article content from its newspaper ‘vehicle’.

The newspaper as we have known it has been shredded.  This is not the fault of Google or any other mechanical process, but rather is a natural if unforeseen consequence of the nature of hypertext.  We are the ones who feel the lure of the link; no machine can do that.  Newspapers made the brave decision to situate themselves as islands within a sea of hypertext.  Though they might believe themselves singular, they are not the only islands in the sea.  And we all have boats.  That was bad enough, but the islands themselves are dissolving, leaving nothing behind but metaphorical clots of dirt in murky water.

The lure of the link has a two-fold effect on our behavior.  With its centrifugal force, it is constantly pulling us away from wherever we are.  It also presents us with an opportunity cost.  When we load that 10,000-word essay from the New York Times Magazine into our browser window, we’re making a conscious decision to dedicate time and effort to digesting that article. That’s a big commitment.  If we’re lucky – if there are no emergencies or calls on the mobile or other interruptions – we’ll finish it.  Otherwise, it might stay open in a browser tab for days, silently pleading for completion or closure. Every time we come across something substantial, something lengthy and dense, we run an internal calculation: Do I have time for this?  Does my need and interest outweigh all of the other demands upon my attention?  Can I focus?

In most circumstances, we will decline the challenge.  Whatever it is, it is not salient enough, not alluring enough.  It is not so much that we fear commitment as we feel the pressing weight of our other commitments.  We have other places to spend our limited attention.  This calculation and decision has recently been codified into an acronym: “tl;dr”, for “too long; didn’t read”.  It may be weighty and important and meaningful, but hey, I’ve got to get caught up on my Twitter feed and my blogs.

The emergence of the ‘tl;dr’ phenomenon – which all of us practice without naming it – has led public intellectuals to decry the ever-shortening attention span.  Attention spans are not shortening: ten year-olds will still drop everything to read a nine-hundred page fantasy novel for eight days.   Instead, attention has entered an era of hypercompetitive development.  Twenty years ago only a few media clamored for our attention.  Now, everything from video games to chatroulette to real-time Twitter feeds to text messages demand our attention.  Absence from any one of them comes with a cost, and that burden weighs upon us, subtly but continuously, all figuring into the calculation we make when we decide to go all in or hold back.

The most obvious effect of this hypercompetitive development of attention is the shortening of the text.  Under the tyranny of ‘tl;dr’ three hundred words seems just about the right length: long enough to make a point, but not so long as to invoke any fear of commitment.  More and more, our diet of text comes in these ‘bite-sized’ chunks.  Again, public intellectuals have predicted that this will lead to a dumbing-down of culture, as we lose the depth in everything.  The truth is more complex.  Our diet will continue to consist of a mixture of short and long-form texts.  In truth, we do more reading today than ten years ago, precisely because so much information is being presented to us in short form.  It is digestible.  But it need not be vacuous.  Countless specialty blogs deliver highly-concentrated texts to audiences who need no introduction to the subject material.  They always reference their sources, so that if you want to dive in and read the lengthy source work, you are free to commit.  Here, the phenomenon of ‘tl;dr’ reveals its Achilles’ Heel:  shorter the text, the less invested you are.  You give way more easily to centrifugal force.  You are more likely to navigate away.

There is a cost incurred both for substance and the lack thereof.  Such are the dilemmas of hypertext.

II:  Schwarzschild Radius

It appears inarguable that 2010 is the Year of the Electronic Book.  The stars have finally aligned: there is a critical mass of usable, well-designed technology, broad acceptance (even anticipation) within the public, and an agreement among publishers that revenue models do exist. Amazon and its Kindle (and various software simulators for PCs and smartphones) have proven the existence of a market.  Apple’s recently-released iPad is quintessentially a vehicle for iBooks, its own bookstore-and-book-reader package.  Within a few years, tens of millions of both devices, their clones and close copies will be in the hands of readers throughout the world.  The electronic book is an inevitability.

At this point a question needs to be asked: what’s so electronic about an electronic book?  If I open the Stanza application on my iPhone, and begin reading George Orwell’s Nineteen Eighty-Four, I am presented with something that looks utterly familiar.  Too familiar.  This is not an electronic book.  This is ‘publishing in light’.  I believe it essential that we discriminate between the two, because the same commercial forces which have driven links from online newspapers and magazines will strip the term ‘electronic book’ of all of its meaning.  An electronic book is not simply a one-for-one translation of a typeset text into UTF-8 characters.  It doesn’t even necessarily begin with that translation.  Instead, first consider the text qua text.  What is it?  Who is it speaking to?  What is it speaking about?

These questions are important – essential – if we want to avoid turning living typeset texts into dead texts published in light.  That act of murder would give us less than we had before, because the published in light texts essentially disavow the medium within which they are situated.  They are less useful than typeset texts, purposely stripped of their utility to be shoehorned into a new medium.  This serves the economic purposes of publishers – interested in maximizing revenue while minimizing costs – but does nothing for the reader.  Nor does it make the electronic book an intrinsically alluring object.  That’s an interesting point to consider, because hypertext is intrinsically alluring.  The reason for the phenomenal, all-encompassing growth of the Web from 1994 through 2000 was because it seduced everyone who has any relationship to the text.  If an electronic book does not offer a new relationship to the text, then what precisely is the point?  Portability?  Ubiquity?  These are nice features, to be sure, but they are not, in themselves, overwhelmingly alluring.  This is the visible difference between a book that has been printed in light and an electronic book: the electronic book offers a qualitatively different experience of the text, one which is impossibly alluring.  At its most obvious level, it is the difference between Encyclopedia Britannica and Wikipedia.

Publishers will resist the allure of the electronic book, seeing no reason to change what they do simply to satisfy the demands of a new medium.  But then, we know that monks did not alter the practices within the scriptorium until printed texts had become ubiquitous throughout Europe.  Today’s publishers face a similar obsolescence; unless they adapt their publishing techniques appropriately, they will rapidly be replaced by publishers who choose to embrace the electronic book as a medium,.  For the next five years we will exist in an interregnum, as books published in light make way for true electronic books.

What does the electronic book look like?  Does it differ at all from the hyperdocuments we are familiar with today?  In fifteen years of design experimentation, we’ve learned a lot of ways to present, abstract and play with text.  All of these are immediately applicable to the electronic book.  The electronic book should represent the best of 2010 has to offer and move forward from that point into regions unexplored.  The printed volume took nearly fifty years to evolve into its familiar hand-sized editions.  Before that, the form of the manuscript volume – chained to a desk or placed upon an altar – dictated the size of the book.  We shouldn’t try to constrain our idea of what an electronic book can be based upon what the book has been.  Over the next few years, our innovations will surprise us.  We won’t really know what the electronic book looks like until we’ve had plenty of time to play with them.

The electronic book will not be immune from the centrifugal force which is inherent to the medium.  Every link, every opportunity to depart from the linear inertia of the text, presents the same tension as within any other hyperdocument.  Yet we come to books with a sense of commitment.  We want to finish them.  But what, exactly do we want to finish?  The electronic book must necessarily reveal the interconnectedness of all ideas, of all writings – just as the Web does.  So does an electronic book have a beginning and an end?  Or is it simply a densely clustered set of texts with a well-defined path traversing them?  From the vantage point of 2010 this may seem like a faintly ridiculous question.  I doubt that will be the case in 2020, when perhaps half of our new books are electronic books.  The more that the electronic book yields itself to the medium which constitutes it, the more useful it becomes – and the less like a book.  There is no way that the electronic book can remain apart, indifferent and pure.  It will become a hybrid, fluid thing, without clear beginnings or endings, but rather with a concentration of significance and meaning that rises and falls depending on the needs and intent of the reader.  More of a gradient than a boundary.

It remains unclear how any such construction can constitute an economically successful entity.  Ted Nelson’s “Project Xanadu” anticipated this chaos thirty-five years ago, and provided a solution: ‘transclusion’, which allows hyperdocuments to be referenced and enclosed within other hyperdocuments, ensuring the proper preservation of copyright throughout the hypertext universe.  The Web provides no such mechanism, and although it is possible that one could be hacked into our current models, it seems very unlikely that this will happen.  This is the intuitive fear of the commercial publishers: they see their market dissolving as the sharp edges disappear.  Hence, they tightly grasp their publications and copyrights, publishing in light because it at least presents no slippery slope into financial catastrophe.

We come now to a line which we need to cross very carefully and very consciously, the ‘Schwarzschild Radius’ of electronic books.  (For those not familiar with astrophysics, the Schwarzschild Radius is the boundary to a black hole.  Once you’re on the wrong side you’re doomed to fall all the way in.)  On one side – our side – things look much as they do today.  Books are published in light, the economic model is preserved, and readers enjoy a digital experience which is a facsimile of the physical.  On the other side, electronic books rapidly become almost completely unrecognizable.  It’s not just the financial model which disintegrates.  As everything becomes more densely electrified, more subject to the centrifugal force of the medium, and as we become more familiar with the medium itself, everything begins to deform.  The text, linear for tens or hundreds of thousands of words, fragments into convenient chunks, the shortest of which looks more like a tweet than a paragraph, the longest of which only occasionally runs for more than a thousand words.  Each of these fragments points directly at its antecedent and descendant, or rather at its antecedents and descendants, because it is quite likely that there is more than one of each, simply because there can be more than one of each.  The primacy of the single narrative can not withstand the centrifugal force of the medium, any more than the newspaper or the magazine could.  Texts will present themselves as intense multiplicity, something that is neither a branching narrative nor a straight line, but which possesses elements of both.  This will completely confound our expectations of linearity in the text.

We are today quite used to discontinuous leaps in our texts, though we have not mastered how to maintain our place as we branch ever outward, a fault more of our nervous systems than our browsers.  We have a finite ability to track and backtrack; even with the support of the infinitely patient and infinitely impressionable computer, we lose our way, become distracted, or simply move on.  This is the greatest threat to the book, that it simply expands beyond our ability to focus upon it.  Our consciousness can entertain a universe of thought, but it can not entertain the entire universe at once.  Yet our electronic books, as they thread together and merge within the greater sea of hyperdocuments, will become one with the universe of human thought, eventually becoming inseparable from it.  With no beginning and no ending, just a series of ‘and-and-and’, as the various nodes, strung together by need or desire, assemble upon demand, the entire notion of a book as something discrete, and for that reason, significant, is abandoned, replaced by a unity, a nirvana of the text, where nothing is really separate from anything else.

What ever happened to the book?  It exploded in a paroxysm of joy, dissolved into union with every other human thought, and disappeared forever.  This is not an ending, any more than birth is an ending.  But it is a transition, at least as profound and comprehensive as the invention of moveable type.  It’s our great good luck to live in the midst of this transition, astride the dilemmas of hypertext and the contradictions of the electronic book.  Transitions are chaotic, but they are also fecund.  The seeds of the new grow in the humus of the old.  (And if it all seems sudden and sinister, I’ll simply note that Nietzsche said that new era nearly always looks demonic to the age it obsolesces.)

III:  Finnegans Wiki

So what of Aristotle?  What does this mean for the narrative?  It is easy to conceive of a world where non-fiction texts simply dissolve into the universal sea of texts.  But what about stories?  From time out of mind we have listened to stories told by the campfire.  The Iliad, The Mahabharata, and Beowolf held listeners spellbound as the storyteller wove the tale.  For hours at a time we maintained our attention and focus as the stories that told us who we are and our place in the world traveled down the generations.

Will we lose all of this?  Can narratives stand up against the centrifugal forces of hypertext?  Authors and publishers both seem assured that whatever happens to non-fiction texts, the literary text will remain pure and untouched, even as it becomes a wholly electronic form.  The lure of the literary text is that it takes you on a singular journey, from beginning to end, within the universe of the author’s mind.  There are no distractions, no interruptions, unless the author has expressly put them there in order to add tension to the plot.  A well-written literary text – and even a poorly-written but well-plotted ‘page-turner’ – has the capacity to hold the reader tight within the momentum of linearity. Something is a ‘page-turner’ precisely because its forward momentum effectively blocks the centrifugal force.  We occasionally stay up all night reading a book that we ‘couldn’t put down’, precisely because of this momentum.  It is easy to imagine that every literary text which doesn’t meet this higher standard of seduction will simply fail as an electronic book, unable to counter the overwhelming lure of the medium.

This is something we never encountered with printed books: until the mid-20th century, the only competition for printed books was other printed books.  Now the entire Web – already quite alluring and only growing more so – offers itself up in competition for attention, along with television and films and podcasts and Facebook and Twitter and everything else that has so suddenly become a regular feature of our media diet.  How can any text hope to stand against that?

And yet, some do.  Children unplugged to read each of the increasingly-lengthy Harry Potter novels, as teenagers did for the Twilight series.  Adults regularly buy the latest novel by Dan Brown in numbers that boggle the imagination.  None of this is high literature, but it is literature capable of resisting all our alluring distractions.  This is one path that the book will follow, one way it will stay true to Aristotle and the requirements of the narrative arc.  We will not lose our stories, but it may be that, like blockbuster films, they will become more self-consciously hollow, manipulative, and broad.  That is one direction, a direction literary publishers will pursue, because that’s where the money lies.

There are two other paths open for literature, nearly diametrically opposed.  The first was taken by JRR Tolkien in The Lord of the Rings.  Although hugely popular, the three-book series has never been described as a ‘page-turner’, being too digressive and leisurely, yet, for all that, entirely captivating.  Tolkien imagined a new universe – or rather, retrieved one from the fragments of Northern European mythology – and placed his readers squarely within it.  And although readers do finish the book, in a very real sense they do not leave that universe.  The fantasy genre, which Tolkien single-handedly invented with The Lord of the Rings, sells tens of millions of books every year, and the universe of Middle-earth, the archetypal fantasy world, has become the playground for millions who want to explore their own imaginations.  Tolkien’s magnum opus lends itself to hypertext; it is one of the few literary works to come complete with a set of appendices to deepen the experience of the universe of the books.  Online, the fans of Middle-earth have created seemingly endless resources to explore, explain, and maintain the fantasy.  Middle-earth launches off the page, driven by its own centrifugal force, its own drive to unpack itself into a much broader space, both within the reader’s mind and online, in the collective space of all of the work’s readers.  This is another direction for the book.  While every author will not be a Tolkien, a few authors will work hard to create a universe so potent and broad that readers will be tempted to inhabit it.  (Some argue that this is the secret of JK Rowling’s success.)

Finally, there is another path open for the literary text, one which refuses to ignore the medium that constitutes it, which embraces all of the ambiguity and multiplicity and liminality of hypertext.  There have been numerous attempts at ‘hypertext fiction’; nearly all of them have been unreadable failures.  But there is one text which stands apart, both because it anticipated our current predicament, and because it chose to embrace its contradictions and dilemmas.  The book was written and published before the digital computer had been invented, yet even features an innovation which is reminiscent of hypertext.  That work is James Joyce’s Finnegans Wake, and it was Joyce’s deliberate effort to make each word choice a layered exploration of meaning that gives the text such power.  It should be gibberish, but anyone who has read Finnegans Wake knows it is precisely the opposite.  The text is overloaded with meaning, so much so that the mind can’t take it all in.  Hypertext has been a help; there are a few wikis which attempt to make linkages between the text and its various derived meanings (the maunderings of four generations of graduate students and Joycephiles), and it may even be that – in another twenty years or so – the wikis will begin to encompass much of what Joyce meant.  But there is another possibility.  In so fundamentally overloading the text, implicitly creating a link from every single word to something else, Joyce wanted to point to where we were headed.  In this, Finnegans Wake could be seen as a type of science fiction, not a dystopian critique like Aldous Huxley’s Brave New World, nor the transhumanist apotheosis of Olaf Stapledon’s Star Maker (both near-contemporary works) but rather a text that pointed the way to what all texts would become, performance by example.  As texts become electronic, as they melt and dissolve and  link together densely, meaning multiplies exponentially.  Every sentence, and every word in every sentence, can send you flying in almost any direction.  The tension within this text (there will be only one text) will make reading an exciting, exhilarating, dizzying experience – as it is for those who dedicate themselves to Finnegans Wake.

It has been said that all of human culture could be reconstituted from Finnegans Wake.  As our texts become one, as they become one hyperconnected mass of human expression, that new thing will become synonymous with culture.  Everything will be there, all strung together.  And that’s what happened to the book.