The web: q’est-ce que c’est?

Chris Anderson and Michael Wolff have written in Wired magazine that, in their words, the web is dead. The web, as opposed to the Internet. Michael Wolff was recently on the Brian Lehrer Show on WNYC, our local public radio station, talking about the article. Here’s Brian Lehrer’s introduction:

At its peak around the year 2000, the web accounted for more than 50% of all Internet traffic in the United States. With the explosion of online video and peer-to-peer communication software, that number has shrunk to around 23% of total Internet use, prompting Wired magazine reporters Chris Anderson and Michael Wolff to offer a post-mortem analysis that they call the death of the worldwide web. Their article, entitled The Web is dead. Long Live the Internet, blames the decline on two things, and argues persuasively for each. Anderson blames our own predilections and the inevitable course of capitalism for narrowing our Internet use to things that we can get easily through the world of apps. Wolff blames them: deep-pocketed businessmen seeking to control the web like the media moguls who controlled radio and television.

It seems clear that the introduction takes data from the article, not from some independent sources. In any case, the key, of course, is what defines the web. Mr Lehrer asks Mr Wolff that question right up front, and here’s the author’s response:

Well, actually, this was an internal debate as we wrote this article, too, with Chris saying the web was one thing, and me saying the web was another, so, I think, at the heart of this debate that question is not exactly resolved. Having said that, remember, what we think of the web is actually a browser. We are looking at something through a browser, through a set of protocols, a construct. Whereas, the Internet itself is a pipe. It just moves things. It is the distribution mechanism of our day; it will be the distribution mechanism of the future. Everything, the overwhelming amount of information that we get will continue to run over the Internet. That’s not what we’re arguing. We’re arguing that this thing that we know as web, think of them as web sites, are more and more a thing of the past.

Mr Wolff seems to be saying that the web comprises the things you look at in your browser, which we’re used to calling web sites. Except that he doesn’t include Facebook in his definition of the web; rather, he gives Facebook as a prime example of how we’re moving away from the web, and using the Internet for other things:

Facebook is something that you reach over the web. I think people can think of it as the web. But it actually is something quite separate from the web itself. As a matter of fact, one of the effects of Facebook is to keep people from going to the web. Everybody is satisfied with what they find on... well, not everybody, but many people spend an enormous amount of time on Facebook itself, never venturing out to the web.

At this point, I have to say that I think their thesis is ridiculous. They’re defining things their own way in order to support what they want to say, and it makes what they’re saying meaningless. When we look at the article, we see a bizarre distinction between apps (not the web) and the web right at the beginning of Mr Anderson’s part:

You wake up and check your email on your bedside iPad — that’s one app. During breakfast you browse Facebook, Twitter, and The New York Times — three more apps. On the way to the office, you listen to a podcast on your smartphone. Another app. At work, you scroll through RSS feeds in a reader and have Skype and IM conversations. More apps. At the end of the day, you come home, make dinner while listening to Pandora, play some games on Xbox Live, and watch a movie on Netflix’s streaming service.

Wait: Facebook and Twitter aren’t part of the web? That’s a stretch, but they’re clearly separating social networking out — illegitimately — to make their point. But the New York Times? Not part of the web? It’s an app? Excuse me: that’s crap.

OK, let’s back up and look at a more reasonable definition of the web.

When the worldwide web first came out of CERN (they had the first web site, in 1991), the distinction between the web and everything else (email, FTP, telnet, gopher, and so on) was, indeed, a set of protocols: web sites were based on HTML (HyperText Markup Language), and data was sent around using HTTP (HyperText Transfer Protocol). That’s still how charts like the one at the top of the Wired article measure traffic. When the chart compares web traffic with email, newsgroups, and FTP, it’s looking at how much traffic goes over HTTP, compared with SMTP, POP, IMAP, NNTP, and FTP.

But first, that’s misleading. It’s valid for what it is — a measure of the number of bytes shoved around through each protocol — but it’s of little value in getting a sense of anything. One email message, for example, will be sent over perhaps 6 to 10 SMTP transactions (maybe more), before being retrieved through POP. A 10 kilobyte email message might account for 100 kilobytes of email traffic. What’s more, if you read the email message through a webmail interface, it counts as web traffic as well, even though most of us agree that email doesn’t really count as the web. Some protocols are also chattier than others, and some data formats add a lot of bytes to what one is sending.

Second, since their straightforward start, the web and the browsers — and other programs — that access it have broadened in scope. Webmail is one example. Browsers also have long supported FTP, so are FTP sites part of the web, or not? News feeds, through RSS and Atom, are largely replacing NNTP, and feeds are retrieved using HTTP. That means they count as web traffic in that chart, yet the authors clearly consider RSS feeds as being outside the web. I disagree: I think they’re very much a part of it.

So, let’s drop the idea of defining the web just by the protocols the information goes over or the programs used to look at the information, and instead let’s go back to what Tim Berners-Lee’s team was building when they put this all together. The idea was to establish an interconnected web of information and services, where things could link to other things, and, through those links, users could find what they were looking for (whether they knew they were looking for it, or not).

When someone posts to Facebook a link to a news article (on the NY Times, maybe), both Facebook and the New York Times site are part of the web. When there’s a Facebook app that lets you post to Twitter from Facebook, and your tweet includes a link to someone’s blog about a peer-to-peer-shared music video, all of that — Facebook, Twitter, the blog, and the peer-to-peer network — are part of the web. Cloud computing services, interlinked through and usable by photo services, blogs, and online document composition applications, are also part of it.

It’s silly to say that the web comprises only static web sites that are written in HTML and delivered through HTTP, and that everything else is something else. That view is from nearly 20 years ago, and it’s long obsolete. It’s obvious that protocols, applications, browsers, and uses will grow and change over time, and it’s natural that they should.

If the authors want to claim that that means the web is dead, well, they can do that... but where’s the value in it? It’s not that anything has died in any real sense, and there’s no one to blame for it. Internet technology is changing in ways that we’ve always expected it would, and in ways we never imagined 20 years ago. That’s a good thing, not a death, and not anything we should establish culpability for.

---

A side note: Michael Wolff comes across as arrogant and annoying on the radio show. For example, when Mr Lehrer quotes a city/suburbs analogy that Virginia Heffernan makes in the New York Times; Mr Wolff rudely interrupts, saying, I always distrust whatever the New York Times says about this because they know nothing about it. That’s why they’re at the New York Times. (Mr Lehrer responds, Well, maybe Virginia knows something, but go ahead.)

The arrogance doesn’t, in itself, invalidate what he has to say. Yet I can’t help but think that it’s part of what’s driven the authors to define things exactly to support what they’re saying.