Dip into the comments, the American text genre of the moment, and you can sense the currents that move American life. “The comments are where the real America is,” someone once wrote in a comment thread. Comments also may be the most obnoxious development of the Web, the wild back alleys where people sound their acid yawps. Their tenor is now so bad that last November a game designer, Shane Liesegang, set up a Twitter feed, @avoidcomments, where he tweets things like “I once showed a comments section to a man in Reno, just to watch him cry.” He now has 27,000 followers.
When we complain about comments, I've noticed, we do so as if we're dealing with some emanation of human nature or the lusty democratic energies of the American soul. But when I went digging into the history of the Web to find out where online comments really came from, it's clear that they're the consequences of what was technically feasible at a certain point and how that feasibility was subsequently implemented. We tend to think that comments represent the culture, but in fact the distinct culture of commenting grew out of digital constraints. Given what Web users had to work with, comments were bound to get weedy.
Comments as we know them — lines of text stacked atop one another in chronological order — are direct descendants of bulletin-board systems, or B.B.S., which date to the 1970s; users could dial in with a modem and contribute to discussion forums. The computer code that determined the order in which text appears on a B.B.S. also provides the basic architecture of the comment thread. That code, or script, became the basis for an early commenting function called the “guest book”: a place for simple text entry in which any visitor could type a note. Guest books were attached to the Web site as a whole, not to any specific content on it. This created confusion about the sort of opportunity that the guest book presented. Was it the soapbox of the online world? Or was it a bathroom wall?
Fray.com, a popular Web site started in 1996, which used freeware guest-book script, was among the first to harness the potential of the guest book. A question would be posed, and answers contributed by readers would be published in real time. The Web site was essentially a collection of guest books, each one with a prompt, and the replies were all stand-alone contributions, adrift in a sea of other people who were speaking to themselves.
It's difficult to give credit for the invention of comments as we now understand them to one specific person or site. An interactive book, “Travels With Samantha,” which allowed readers to submit comments via a form (much as modern comments are), won a Best of the Web award for document design in 1994, which was the same year that a consortium of World Wide Web developers created W3 Interactive Talk to discuss technical matters — on which discussion points were submitted by a form that made them part of a topic page.
There are competing claims about the first blog to offer comments. One pioneer is Bruce Ableson, who created Open Diary in 1998, an Internet journaling project that allowed diarists to respond to one another's entries. Around the same time, the software developer Dave Winer created a discussion board with a commenting feature. The first comment there arrived on Oct. 5, 1998: “Too bad coders can't be like rock stars and get their money for nothing and their chicks for free.”
As the Web evolved, the simple top-to-bottom arrangement of autonomous lines of text had a huge influence on the culture of commenting — and it now feels as natural as having sidewalks on two sides of the street. But the typical Web page presents a hierarchy. At the top is the article, post, video or image, while the comments dangle underneath like s**erfish from an indifferent shark.
“Having the comments at the bottom of the page — people feel that; they feel they're not as legitimate a voice as the original post,” says Travis Nichols, who moderated a Poetry Foundation blog for several years and drew on this experience to write a novel, “The More You Ignore Me,” whose belligerent narrator takes over a cooking Web site with a venomous lament. Commenters show up with an immediate grudge, Nichols observed — they know they've been relegated to the steerage cla** of the public discourse.
As a result, commenters became territorial. “When we had to be more aggressive in deleting comments that violated our commenting policy,” says Bill Adee, The Chicago Tribune's vice president of digital operations, “I got far more complaints about deleting comments than I did about the level of discourse.”
In the late '90s, early bloggers expected the level of discourse to be high. In fact, intelligent commenting was seen as a path to gaining respect in the blogging community. At the beginning of 1999, there were only about two dozen blogs (which were mainly lists of interesting Web sites), but as the number exploded, it became hard for bloggers to follow the fragmenting conversations. In 2000, the blog service Blogger introduced permalinks, which allowed each blog entry to have its own URL, and in 2002, Moveable Type implemented the TrackBack, which automatically alerted an author that a permalink from his blog had been posted elsewhere. The TrackBack was meant, at least in part, to blur the lines between commenters and writers; the conversation surrounding one blog post no longer needed to be relegated to the comments section, but could be sprinkled across disparate blogs with the TrackBack as its link. That was great, in theory. But while conversations were the model for interactions, the technology couldn't sustain what real conversations required.
What k**ed the promise of the culture of openness in early blog culture was the blizzard of link spam that hit in the mid-2000s. Planting millions of links to penis-enlargement ads in the comments of the most innocuous of blogs could trick Google bots into ranking certain penis-enlargement services higher when actual humans search “penis enlargement” (which, apparently, they do). Cleaning spam out of comment sections became a bigger headache for bloggers than policing the trolls. A lot of them disabled comments for good.
The invasion of spam highlighted the problems with early blog moderation, which grew out of a culture shaped by attitudes and tools inherited from the free-speech cowboys of the B.B.S. era. It took site owners years to realize that they're not providing platforms and soapboxes but creating communities out of lines of text, which requires a more subtle approach. The Web forum MetaFilter, for instance, which is known for a positive commenting flavor, depends on a 24/7 team of moderators. “People come to us all the time and say, ‘Here's a problem with people behaving badly, we want a tech solution,' ” says Paul Bausch, a MetaFilter developer. “We tell them that human problems require human judgment.”
Yet high-traffic sites continue to leave comments unmoderated or use imperfect automated moderation. Only a few seem to have tried user-moderation systems like the one developed by Slashdot's creator, Rob Malda. Founded in 1997, Slashdot rapidly began to suffer from what Malda called “signal-to-noise-ratio problems” as tens of thousands of users showed up. Rather than embracing the chaos (which was a hallmark of Usenet, another digital channel of communications) or locking things down with moderators (which e-mail lists did), Malda figured out a way for users to moderate one another. Moderation became like jury duty, something you were called to do.
In my view, the worst places to visit aren't the comment jungles of 4chan or YouTube, but the overly manicured comment lawns of some newspapers. Papers have mistakenly treated comments as the digital equivalents of letters to the editor. “We've got a 160-year tradition of no comments on our stories in the newspaper, so it's not surprising it took a little bit of time to get comfortable with that idea,” Adee of The Chicago Tribune says. (At The Times, select articles, including this one, are open for comments, which are moderated by humans.)
Talking to people at newspapers makes it seem as if the future of comments is all social log-ins and filtering algorithms. But these are really just tools for putting a lid on commenting culture's excesses, not rethinking the relationship between creators and commenters in more fundamental ways.
A step in that direction is annotation, in which reactions, corrections and elaborations are placed directly on the text itself, which could, perhaps dangerously, put commenters on the same plane as writers and reporters, who spend days or weeks or months learning about a subject. One example is Medium, which allows readers to make notes at the paragraph level. (Unlike a Wikipedia entry, where users can edit the text, the article remains intact.) Writers might balk at this, but look at it this way: people are more likely to comment on what's in the text, which may prompt them to actually read it before commenting.
Another is the start-up Rap Genius, a community built initially around annotating rap lyrics. (They've expanded to things like legal decisions and poetry.) The premise is that keeping readers and commenters close to the text focuses activity and keeps the discourse informed and civil. There are still people who act like idiots, but much of the time, it's because they don't know what the community is trying to achieve. The goal, Rap Genius's co-founder Tom Lehman said, is to help out people who are making real contributions by defining and rewarding “real contributions” with user badges like “Rap I.Q.”
What makes Rap Genius an especially big deal is that it received a $15 million investment from a venture-capital firm run partly by Marc Andreessen, one of the authors of Mosaic, the first widely adopted Web browser. It turns out that annotation was an original feature in Mosaic, but it was dropped in part because the annotations would have had to have been hosted on the company's own server, which was too expensive. Plus, Andreessen said, content creators (understandably) weren't crazy about letting “Joe random user” annotate their stuff. Interestingly, Andreessen said he had been reading a lot of postmodern literary theory at that time, so the relationship of annotations to texts was “one of the threads I was pulling on.”
It's a fascinating alternative-history proposition: would a world of annotations, rather than comments, inspired in part by Jacques Derrida, have set the Web on a different course? Social media might look very different; you can easily imagine an alternate version of Facebook and Twitter made up of people who regularly annotate certain sites across the Web. In this version of the Web, people would be writing on the worldwide wall of history, not scrawling in little spaces under siloed bits of content. But maybe there, too, they'd never figure out what they were supposed to be writing together, except to a**ert that everything they individually believe is true. America, Allen Ginsberg might have said in his sarcastic poem of the same name: Your caps lock is on.