How Google accidentally built a truly beloved social network, only to steamroll it with Google+ . The sad, surprising story of Google Reader.
Last October, while hundreds of protesters were encamped in Zuccotti Park, a handful of people occupied a glass building in downtown Washington D.C. Wearing sheepish grins and business casual attire, the ninety-nine percent they were not; one demonstrator said he worked for Grover Norquist. “GOOGLE: DON’T MARK ALL AS READ,” pled a poster outside the Internet company’s D.C. headquarters, where the aggrieved customers had assembled.
That week, on the heels of launching its long-awaited social network, Google was to make significant—and feature-breaking—changes to Reader, an RSS news aggregator launched in 2005. Formulated as a “river of news,” a scrollable collection of headlines from across the web, the site added sharing features that endeared it to a niche group of users.
Pre-Twitter, it was the essential aggregation tool for news and information junkies. But Reader had also became a social network in its own right. Four years on, with Google+ ascendent, these same social functions were marked for elimination. And so, its users fretted, was their beloved Google Reader.
“We think the end result is better than what’s available today,” wrote Alan Green, a software engineer who announced the impending changes on October 20, 2011. “We recognize, however, that some of you may feel like the product is no longer for you.” For Reader loyalists the revamp was not only seen as worse—it was devastating. Over the years, so-called “Reader Parties” had sprung up everywhere from the Long Island Sound to the Ozarks.
“Sharebros,” as some came to be known, crashed on each other’s couches, sparked real-life friendships, and not a few marriages; in the soon-to-be-appropriated language of Google+, there were “hangouts” among “circles.” Google had fostered a social network and earned die-hard fans in the most valuable way possible — without trying.
The few picketers in Washington stood for numerous members in a Facebook group that spanned from Houston to Detroit, with outposts in Macedonia, Switzerland and Singapore, and a particularly vocal contingent in Tehran. After an Iranian blogger lobbied against the changes (citing Reader’s contributions to the Green Movement), a barrage of press from TechCrunch, Mashable and Andrew Sullivan directed 13,745 signatures to an online petition. Nonetheless, on Halloween of last year, Google went ahead with the redesign. A former product manager called it a disaster, while an ex-designer offered to rejoin the company to stanch the damage. As one commentator complained, “There’s almost no way Google could have blundered more disastrously here.”
In the year since, Plus has been derided as a “virtual ghost town,” and a “complete failure” unpopular even with Google employees. All of which has heightened the resentment shared by Reader fanatics. Today, they are a population dispossessed. Many have disappeared off the grid, while others struggle to rebuild communities that were, with a few keystrokes, deleted. All of them, the dental student in San Antonio, the academic librarian in Boston, the game developer in San Francisco, yearn for the scroll-tracked Shangri-La that was.
They wonder why Google deep-sixed superlative features, years in the making, for an upstart social network, a Facebook clone. In the year past, the same question has been framed and phrased in a thousand different ways—why force an unproven social network on users at the expense of an organic one?
Or, to put it more bluntly, why pave over paradise to put up a parking lot?
***
In the summer of 1999, Meg Hourihan took a vacation. She had recently founded Pyra Labs, a Silicon Valley tech startup that aimed to increase workplace productivity. Hourihan was also the author of Megnut.com, one of the earliest weblogs. While she was away, her business partner, Evan Williams, who went on to co-found Twitter, released a side project, a digital publishing service for just such budding diarists. The project found an audience, and soon became Pyra’s focus. Blogging was taking off, although the Pyra founders said they worried about “unimaginative” postings like “I’m tired” or “This sucks.” The following spring, Hourihan met Jason Kottke, a web designer who, among other creations, drew the “half-assed technicolor logo” on Gawker’s masthead. Kottke was another blogging pioneer, and after a courtship online as much as off, he and Hourihan married. (Disclosure: Kottke sometimes works from a desk in the BuzzFeed offices.)
In early 2001, Chris Wetherell, a software engineer, was reading Kottke’s blog. That day, a company called Moreover was featured. Moreover was offering what is today a hallmark of any website’s right rail: a feed of headlines from around the web. To find the desired stories, you entered keywords: “Britney Spears, Napster, DVDs, museums, etc.” for example. Wetherell wanted to try it himself. His version, called Java Collect, was an open-source library that displayed Moreover’s feeds and scraped websites like CNN for top news; on 9/11, when news sites went down, Wetherell’s module retained some of their cached headlines. Over the next few years, while his servers spun at home, Wetherell joined Google. He was working on Blogger, the service that Hourihan and Williams had started and, in 2003, sold to the burgeoning, and still privately-held, search company.
Blogger, too, displayed feeds on their pages. These feeds, also known as web feeds or news feeds, first came about in the late 1990s. At that time, early bloggers like Williams and Dave Winer, a software developer, had made meaningful headway in publication, but struggled with syndication. Blogs would be updated without their readers ever noticing. While the web had embraced HTML, much of its sister language, Extensible Markup Language, or XML, had yet to be written. If HTML exists to display data, XML was built to transport it. In XML, Winer saw a solution to the syndication problem. And in April of 1997, he released Scripting News, a sort of “bibliography for DaveNet,” his blog at the time. “I decided to take the plunge, and take the chicken out of the chicken and egg problem. Or is it the egg? Who knows. But from now on, Scripting News, in addition to being a HTML web page, is also an XML application,” Winer wrote. With the advent of web feeds, blogs like DaveNet, EvHead or Megnut could reach beyond a dedicated audience to find wide distribution. The web’s Gutenberg moment had passed, and here were intimations of a fax machine.
Called RSS (for Rich Site Summary, and later, Really Simple Syndication) the new technology followed a volatile and often contentious trajectory over the next five years. Engineers at Netscape designed their own version for My Netscape, a forerunner of personalized web portals like iGoogle. But when Netscape was acquired by America Online, development on RSS grounded to a halt. And that’s when the infighting started. Winer took up a post at Harvard’s Berkman Center, where he released RSS 2.0 under a creative commons license, and tried (but failed) to trademark the name. Meanwhile, developers from IBM, Moreover, and Google’s Blogger developed an alternative version called RDF, motivated as much by technical disputes as by Winer’s alleged imperiousness. “Dave Winer has done a tremendous amount of work on RSS and invented important parts of it and deserves a huge amount of credit for getting us as far as we can. However, just looking around, I observe that there are many people and organizations who seem unable to maintain a good working relationship with Dave,” Tim Bray, one of the creators of XML, blogged in 2003. In some circles, he added, RSS had come to stand for “Reliably Spiteful Squabbling.” Winer fired back: “Why has my personality become the issue? They’re using that to try to get me to shut up.” While haggling with his detractors, Winer cannily sidestepped the dispute, and took his offering straight to the publishers. In April of 2002, the New York Times signed on with RSS, followed by Salon and Rolling Stone . By June of 2004, the Times had expanded its offerings to 27 feeds. No longer an obscure directory for bloggers, as the paper of record reported, “this increasingly popular online tool turns a morass of disparate information sources into an automatically generated and neatly organized index of the latest articles and postings.”
It was later that year when a Google colleague challenged Wetherell to construct an Atom parser in Javascript. Atom was yet another XML language, and the dare called on Wetherell to expand on his old passion project. If successful, he would have “something that turns something into something else which could be used to represent data that was basically about cat photos,” as he transliterated for the layman. The parser worked. One night, while testing and debugging, Wetherell had a Frankenstein moment. “A little wheel reinvention occurred,” he recalled, “as a square. The parser became a reader by accident.” Most of the time, software developers write additional code to describe complex data. By a stroke of luck, Wetherell’s reader actually bypassed a layer of complexity. Wetherell designated the creation as his 20 percent project. In a pitch meeting, he drew a circle on a whiteboard, and wrote below it: “Feed reading is inherently polymorphic.” The kind of RSS reader Wetherell envisioned would be “athletically flexible to match a wide variety of reading styles.” He drew spokes out from the circle to delineate the different use cases he imagined. Finally, his colleagues said, “Ok, then.” Work on “Fusion”—the prototype for what became Google Reader—would begin.
***
A supposed principle of contemporary web behavior is the 90-9-1 rule. For every 100 users, it posits, 90 people will consume content curated by nine influencers, originated by just one creator. It takes a single finger-biting baby and about a million upvotes to generate nearly half a billion views on YouTube. In a more traditional context, a reporter like Michael Hastings gets Stanley McChrystal on the record trashing President Obama, and Politico, The Huffington Post, and Gawker disseminate the story to a news-hungry public. Over the past decade, the ninety’s mindshare has seesawed between the one and the nine; in the McChrystal case, Rolling Stone, for whom Hastings worked at the time (he now writes, additionally, for BuzzFeed), was up in arms when blogs shared advanced copies of its scoop. The relationship between originators and distributors has been described as parasitic (see Keller on Huffington: “flocks of media oxpeckers who ride the backs of pachyderms”); commensalistic (see Zuckerberg to publishers: “discover more news stories through your friends”) and symbiotic (see Peretti on Zuckerberg: “They own the railroad tracks, we drive the trains”). Facebook, Twitter and Tumblr, being mass-distribution platforms for user-generated content, scratch both itches, and hence are touted as the ne plus ultra of software development.
In the beginning, Google Reader was merely a means of distribution—a tool for the 90 percent. The small team, which by 2007 had grown to nine engineers, innovated mostly on improvements to the reading experience, adding labels, search, and embedded videos. The beauty of Reader was that it was self-contained. Fossicking through search engine results meant more tabs, more browser windows, more clicking out; with Reader, it was about digging in. Infoholics used Reader as a one-stop shop for their favorite sites and blogs. There did exist a rudimentary method of sharing, a workaround where one user could subscribe to another’s feed of shared items. But as Christine Eslao, a Reader from Boston, remembers, “you were kind of sharing in a void for no discernible purpose.” Chrix Finne, a Reader product manager, described the sharer’s plight in his limerick:
There once was a Reader named Chrix,
Who found himself in quite a fix:
He'd fun stuff to share
But no one was there,
Now Reader shows friends when Chrix clicks!
In December of 2007, Reader linked up with Google Talk (the chat feature in Gmail) to display shared streams from friends. Within the context of feed reading, it fomented something of a Neolithic Revolution. Foragers, hitherto gathering headlines on a crude and solitary basis, became farmers, cultivating streams of information for their neighbors. Sharing increased twenty-five percent overnight. At Yale University, a student named Richard Berger (later known as Richard Likebot or Obscure Reference) notified his friends of the changes. “Holy shitballs,” replied Tom Lehman (aka Lemon or just Tom), who would later create the popular and lavishly funded lyrics annotation site Rap Genius. “I fully support this idea. Even if no one else does, please add everything to your shared feed.” Soon after, Reader implemented a bevy of complimentary features like pithy annotations and a bookmarklet to aggregate sites that didn’t support RSS. In March of 2009 came a crucial update, allowing Readers to comment on each other’s shares. Eslao, the Reader from Boston, convinced her boyfriend to sign up.
No comments:
Post a Comment