About a year ago I started seeing a lot of headlines about “deep links.” Developers, I soon discovered, are frantically trying to get mobile apps to work together, so that clicking on a link in one app takes you directly to relevant content in another app. Deep linking means to bore a wormhole-tunnel that hops you directly from a specific spot in one app to a spot in another, no side trip to a browser or a home screen needed.
That’s a fine thing. Today’s mobile software is all app fragments; it’s like Thirty Years’ War-era Germany, a patchwork of principalities with shifting allegiances and frequent skirmishes. Anything that knits it together is helpful.
But as I reviewed the coverage I noticed something a little odd. The idea of a deep link has a much deeper history — but no one was making the connection between the hot new trend in mobile and the one that I remembered from the 1990s.
Beginning in the late ’90s, legal spats arose on the Web over “deep links,” defined as links that bypass a site’s home page, whisking you straight to some specific location or product or piece of content. Some businesspeople didn’t like this practice; they thought they were being ripped off by people trying to steal their pageviews. They didn’t understand how the Web works, and every so often they’d push a lawsuit, and lose.
With mobile’s deep links, everyone seems to have collectively wiped the buzzword slate clean and started fresh — no context, no memory. No depth. It’s as if someone started a new comedy act today and called it “Monty Python” without offering any sign of knowing the name had a history.
This cluelessness is extra-ironic because, originally, the exact purpose of links was to make this kind of connection clear. The people who invented the link saw it as a tool for relating ideas in illuminating ways — for making conceptual leaps and connecting disparate thoughts. If these visionaries had achieved their aim, the kind of tech-cultural amnesia represented by the recycling of the term “deep links” shouldn’t have been possible, two decades into the Web era. The links with true depth that they envisioned would have made sure of that.
Today, though, Web links are mostly navigation and footnotes. Instead of sharing linked trails of knowledge that we’ve blazed, we leave piles of data around that service providers mine for value.
How did we end up here? Where are the links that should remind us that “deep links,” as a concept, already exists? Let’s start at the beginning of the story.
How the Web Was Linked
The link as we know it on today’s Web comes down to us from a long tradition of deep-thinking people.
First there was American engineer Vannevar Bush, who wrote a famous 1945 essay in The Atlantic imagining a device called the Memex that researchers could use to forge and capture “associative trails” of thought. Technically speaking, those were different times, and the Memex seemed to call for an awful lot of microfilm. It remained vaporware. But Bush’s big idea — that our technology could and should serve us by putting us in the informational driver’s seat and letting us replay and share our thought-patterns — had legs.
The process of tying two items together is the important thing… Thereafter, at any time, when one of these items is in view, the other can be instantly recalled merely by tapping a button… Moreover, when numerous items have been thus joined together to form a trail, they can be reviewed in turn, rapidly or slowly… Any item can be joined into numerous trails… And [the] trails do not fade.
Two decades later, computer pioneer Douglas Engelbart actually built and demoed a tool that not only achieved Bush’s vision but also made it collaborative. You could click on words in your documents and your screen would reload, and voila, you were at a different place in your document, or another one from your files, or — this was the coolest part — one that a whole group of people had written.
Around the same time, information technologist Ted Nelson coined the word “hypertext” to describe what Engelbart’s system was enabling. Later, experiments such as Apple’s Hypercard and CD-ROM publishing gave a wider population a taste of how that might work. Finally, English computer scientist Tim Berners-Lee took a limited version of this vision and plunked it down on top of the Internet’s networking protocols — and the World Wide Web was born.
Here’s the hardest thing to remember about discovering links at the dawn of the Web: They were fun. As journalist Gary Wolf put it in the lead of a 1994 Wired piece that introduced the Web browser Mosaic to a wide readership: “Mosaic is not the most direct way to find online information. Nor is it the most powerful. It is merely the most pleasurable way… By following the links — click, and the linked document appears — you can travel through the online world along paths of whim and intuition.”
The fun was not simply in the Oz-like, click-and-you’re-somewhere-else delight of moving from link to link. Links were also a joy because they were absurdly simple to make on your own pages, to add new hops to the network. It helped to be a geek, and to be patient, but you didn’t have to be a computer scientist or a programmer to make the magic happen. The puckish creators of Suck.com found another kind of fun in using links as an expressive device: “They buried their links mid-sentence, like riddles, like clues,” Steven Johnson wrote in his 1999 book Interface Culture. “You had to trek out after them to make the sentence cohere.” Their link style only survives in a handful of places today, but it introduced a whole generation of Web creators to the notion that HTML form could convey complex meaning.
As the Web took off we began to understand that links were more than just fun — they had power. They let us remix everything we encountered online. You could use them to bypass the official version of a news front page or a storefront or a government site and erect your own alternative instead. The authors of the 1999 Cluetrain Manifesto — a Web page-turned-book arguing that the Internet was transforming markets, and everything else, into “conversations” — put it succinctly: “Hyperlinks subvert hierarchy.” For a while, links did just that — which was why you’d periodically see companies trying to protect their hierarchies by taking “deep linkers” to court.
Then Google came along and showed us that links could be read as signals of authority and value. It assembled those signals into an index of the Web of unsurpassed quality. Eventually, it found a way to make a fortune by putting a price tag on those links — and soon we were beset by link spam and spam blogs and bogus link farms. Google didn’t mean to ruin links for the rest of us, but that’s what happened. Before Google made it possible to earn microcents by putting text ads on search-optimized pages, a link actually meant that some human being thought to place it on a page because it mattered in some way. Now, the link was a move in a multi-dimensional chess game in which the winner got to collect tiny sums, over and over.
Links suddenly weren’t so much fun any more. They stopped serving us as an alternative way of thinking about and creating informational relationships; they settled into a functional role. They became tools for navigating websites and pointers for sharing content on social networks. Finally, links became click-bait — transparent come-ons for traffic in an accelerating race to the bottom of our brainstems. We found ourselves arguing whether links help us see connections or just distract us or make us stupid.
Today, links are most widely used to reference source material, not to connect ideas. We use links to stock basements with footnotes rather than build cathedrals of context. Most of us think of a link as something to click on or pass around to our friends, but not so much as something to make. And when the tech industry decides that it’s going to create something called “deep links,” its own system of information management and retrieval doesn’t help it see that the phrase already has a record.
Knitting Mobile Together
The Web was a buzzing hive of links from the day it came into wide use. (Occasionally, even tech journalists seem to forget this.) Links are what makes the Web a web; without them, it is unimaginable. They are how it grew, how early users found one another, how they shared stuff and built out their world. The popup ads, the retargeted messages, the monetized links — that all came later.
With mobile, it looks like we’re going to see what happens when you flip this sequence of events — when you try to build an interconnected network of links by starting with the marketing stuff and then, later, try to add the contributions of users.
Chris Maddern is cofounder of Button, one of several companies that have set out to make deep links work in the land of apps, and he talks with rapid precision about the sorry state of mobile interoperability today.
“Right now it’s no secret that the Internet’s paid for basically by big companies buying tiny time-slices of your eyeballs against your will,” Maddern says. Button wants to change that by “capturing users’ intent.” For instance, you’re reading a New York Times travel story about Barcelona. You want to book an Airbnb there pronto. On your phone, you’d have to exit your New York Times app, then start up your Airbnb app and search for Barcelona in it. In a Web browser, you could have clicked straight through from one site to the other — and landed directly on a page of Barcelona listings.
If we can recreate this kind of experience in the app universe — and Button has built a popular open-source kit for developers to do just that — Maddern believes we can replace the old advertising regime with “cascades of affiliate revenue coming back down the network.” Once that happens, it’s only a matter of time before the mobile platforms begin to coalesce into a universe of content and functionality that we can traverse and curate at will — kind of like the Web is already.
“We’re still relying right now on the Web for search and discovery,” Maddern says. “The app-to-app aspect hasn’t really matured yet. Once we get past this craze of, like, ‘omigod, you can deep link things from one app to another,’ we’ll start to think of actually interesting ways of building an index — and being able to discover and drive straight down to both user-generated and brand-generated content inside apps.”
Maddern is pretty confident that this will all happen. Right now, though, we could be forgiven for worrying that deep linking is mostly about helping people sell us stuff. Deeplink.me is another outfit that’s building the mobile deep-link world, and here’s how it describes the experience (in a section of its site helpfully labeled “For Marketers’ Eyes Only”):
User has your app installed, almost buys BLUE SHOES, but leaves app
We create unique, dynamic creative featuring BLUE SHOES
We serve the unique creative to user while they are in other apps
User clicks, and app opens right to cart with BLUE SHOES already added
Great! Now those ads that chase us around the Web reminding us that we once briefly considered purchasing some navy footwear as a gift but thought better of it — those ads can, like digital harpies with dimension-crossing powers, hound us from app to app.
As if being pursued through all your apps by blue shoes wasn’t sinister enough, this current iteration of deep linking has an even darker side. Most apps are proprietary; you can only create deep links to them if the app’s creator allows you to. Without some radical shift in the mobile computing platform, in other words, the hierarchy will not be subverted there.
Meanwhile, back on the Web, the bulk of the links we encounter are lazy, manipulative, or mundane. Is it time to play “Taps” over the link’s corpse? I asked David Weinberger, the Cluetrain co-author who has written beautifully over the years about the Web’s meaning.
“The existence of links as a network of semantic relations is an awesome development in human history, and I don’t think we’re going to lose that,” he said. “I do worry that ordinary people without big design and production companies behind them will forget that they are able to do this as well. At which point links still have all the old qualities — except the sense that they are ours. And that would be a big loss.”
Taking Back Our Thought-Trails
Not everyone has run up a white flag and abandoned claim to those networks of semantic relations. Vannevar Bush’s broader vision — of technology that lets us map and share the “associative trails” of our thoughts and ideas — lives on, however imperfect the tools and limited the results. Deep-linkers of this ilk are enthusiasts of the software niche known as “personal information managers” — the kind of tools that Howard Rheingold used to call “brain amplifiers.” (His book Tools for Thought remains a prescient chronicle of this field from its earliest days.)
This is the genre of mind-mapping, outlining, or note-taking tools that have been with us all the way back to the era of Lotus Agenda, More and Ecco. Every time a writer or speaker creates a project by laying out ideas in a program like Tinderbox, DevonThink, Scrivener, Workflowy, Evernote, or [your favorite here!], she is living out a little of Bush’s Memex dream.
Here’s one example: Since the ‘90s, Jerry Michalski, the longtime tech pundit and consultant, has been building a web of interconnected ideas and links using a mind-mapper called The Brain. “I’ve been curating a big ball of twine of links to everything that matters to me,” is how he describes it to me. It crossed the 200,000 entry mark some time ago, and Michalski adds new links to it every day. It’s public on the Web, and recently became available in iOS app form, too.
Michalski has recorded his thought-trails for his own purposes, but he has also gone to the trouble of sharing them with the world. If, like me, you love exploring different ways to organize and visualize ideas and connections, you can happily get lost in his brain, and indeed in any of the other programs that enable similar experiments. (Steven Johnson has written at length about his use of DevonThink, and The Atlantic’s James Fallows has carefully chronicled this realm for years.) But aside from rare efforts like Michalski’s, mostly these projects produce isolated thought-islands that are difficult to visit and explore. The tools produce material in closed formats that can’t be readily distributed or shared.
Here’s another example that’s all about the collaboration. A few years ago Ward Cunningham, the inventor of the wiki, pressed the reset button on his original concept and emerged with a new idea called “federated wiki,” a kind of cross between the wiki and Github. Old-school wikis, of which Wikipedia is the biggest and best-known, famously let anyone edit each article. With a federated wiki, you only edit your own articles, but you can take anyone else’s and “fork” it — incorporate a version of it that you’re then free to change. These forked texts carry their histories with them; every piece of data has a pedigree. The page’s original owner can incorporate your edits or ignore them.
Today, the federated wiki scene is very much an early-adopters-only kind of club; it’s tough to explain and an uphill climb to participate in. But the concept is gaining some traction among the adventurous. All it needs to take off is to be embraced by one passionate group of users (the way, say, that Grateful Dead fans powered the early evolution of the pioneering online community The Well). A fully realized version of Cunningham’s wiki confederation would be a web of ideas where everyone gets credit for his own contributions but anyone can borrow them, build on them, or contest them — and all the connections remain intact.
Tools like Cunningham’s and Michalski’s demand that we actively tend our “gardens of hypertext” (to borrow a phrase from Tinderbox creator Mark Bernstein). That takes work, work that a lot of us are never going to do. Yet there’s another kind of idea-trail that requires no effort at all.
All That We Leave Behind
Passively, thoughtlessly, each of us could be building Bush’s Memex every day — with every click and swipe on our devices. The links we follow, the pages we read, the photos we tap and the messages we send tell the story of our lives and the thoughts that fill our heads. But this information profile — the record of our use of the Web and the cloud — is mostly out of our reach. It’s available to the companies that build and run our services to analyze and mine and sell stuff to us, but it is rarely similarly available for us to access and share for our own purposes.
A few brave startups and projects aim to redress this imbalance, to put Big Data to work for the little guy. For instance, ThinkUp, from Gina Trapani and Anil Dash, will mine your Twitter and Facebook activity for insights and patterns — then deliver the information to you, not to a marketer.
For now, ThinkUp is an outlier. It’s not that the problem of personal data autonomy isn’t well-known and widely complained about. But, mostly, it’s understood from a few familiar perspectives: Our privacy is being invaded — keep out! Or, the economic value of my data is being stolen — why can’t I cash in?
Meanwhile, there’s another kind of loss happening that we’re less in touch with: a cultural erasure, a mass abandonment of intellectual value, which takes place as our data contrails, their monetizable details extracted, vaporize into the ether. The contexts we assemble in our online existences each day never get a chance to enlighten us over time; instead, we throw them away, or hand them over to Facebook or Google, Apple or Amazon.
Here’s how Michalski puts it: “It’s like we’re leaving a ticker tape of neurons that fired at different points in different ways, and yet from that ticker tape you cannot reconstruct the brain and all the relationships it knows and holds. And nobody seems concerned. We’re all like, ‘Yeah, there’s tons of ticker tape around, look at this — we can hold a parade!’ And we’re drowning in it. There are so few tools that allow us to make context that matters.”
That situation won’t change quickly. It won’t change at all if the only kind of “deep link” we find in our apps is the kind that connects us with stuff to buy and services to employ.
Still: In the ‘90s, we got tired of systems like Compuserve, AOL and Prodigy that wouldn’t play together nicely and only let us play in pre-approved ways. We might similarly grow disenchanted with apps that don’t connect easily, or only connect in ways that we can’t shape.
And then someone might come along with a better system, one that uses links to save and share our thoughts in meaningful, lasting ways — whether we’re using a phone, a pad, a laptop, or a watch.
Make it fun, too, and it can’t miss.
If you want to follow the thought-trails that led to this article, you can. I’ve taken the outline I used while I worked on it (in the outliner app Workflowy) and made it public (with some minor cleanup). These are, essentially, my notes, but since my thinking happens in outline form, you can also get a sense of how these ideas came together — and you’ll find a bunch that didn’t make it into the final cut, too.