The Web Is Dead? A Debate

O’Reilly to Anderson:
It’s the back end that matters.

While there’s no question that both Facebook and the mobile app ecosystem provide clear challenges to “the web,” the idea that the browser front end was ever the key to the web’s dominance is so, well, 1995, from the days when Netscape thought that the “webtop” would displace the desktop. But the competitive action has always been on the internet as transport, with data-driven services as the back end.

Back when I put on my first conference, the Perl Conference, in 1997, I was already talking about how the internet was becoming a vast repository of programmable services, that screen scraping and overloaded URLs were pointing towards a future internet operating system. And when I put on my “Building the Internet Operating System” conference in 2002, I was already focusing on how Peer-to-Peer distribution, distributed computation, and web services were pointing forward to something much bigger than we’d seen before.

And in 2004, when we rechristened this whole thing “Web 2.0,” I was very clear that “Data is the Intel Inside,” that what we were talking about was an internet operating system, whose subsystems were data systems like identity, location, payment, advertising, media repositories, and product ids.

And sure enough, “web sites” like Google, but also now Facebook, Twitter, Amazon, PayPal, LinkedIn and many others, have been quietly building those enormous data back ends that drive their web sites, but more importantly, also drive a vast array of web services. Google maps in the browser is still Google maps, with all the intelligence, all the deep data layers, that make it a success on either front-end.

What the mobile ecosystems of today have done is to unmask the reality that it’s the back end that matters.

The key concept of Web 2.0 was that in an era of networked applications, network effects matter, the rich get richer, and there are enormous concentrations of power inevitable in an era when applications literally get better the more people use them, not just because there are more people using the same application (which is the kind of network effect we saw in the era of Microsoft Word and Excel) but because the applications literally get smarter as they gather more data from their users. (As Google chief scientist Peter Norvig once said about speech recognition, “We don’t have better algorithms than anyone else, we just have more data.”)

So yes, the front end doesn’t matter in quite the way we thought it would. Perhaps a good analogy is to think of the web as the boot loader for the Internet OS – it’s how we got the data up there into the cloud, and got the process of self-reinforcing data collection started. It’s also how we got the business models going that turned all that data into cash.

One other key concept is that it’s the data that provides what John Battelle and I are calling “points of control” in the future Internet operating system. (See Theme: Points of Control) It’s not APIs on the phone, it’s not Objective C or the iPhone OS, it’s still the data back end that gives even Apple its leverage. The reason why iPhone and Android are beating other phones is because they have more apps. It’s really important to see the “app store” (as in the iPhone App Store and the Android Market) as examples of these kinds of massive network-accessible data collections. It’s easy to focus on the apps themselves, down on the phone, and to forget just how many of the key apps are the same networked apps that we see on the web, just with a different front end. The app store itself has more in common with Google Maps than it has with, say, Pocket Pond.

Looking forward to continuing this discussion.

Anderson to O’Reilly:
Open is better, but…

I agree with all of this, but am surprised that you’re not more worried about the consequences of the shift from the “front end” of browser-centric computing to the “back end” of apps, closed networks and proprietary connections between massive data servers and specialized clients.

Why? Because the first defaults to open and the second defaults to closed.

You’re looking at data architectures, and you rightly observe that deep data on servers allows clients to be smaller and lighter, as the app explosion is proving. You call these valuable data sets “points of control” in the emerging Internet Operating System (notably not Web Operating System). And you say they’re earned with data that people want.

All true. But there’s another term for “points of control”: monopolies. From Facebook to iTunes, we are seeing more and more Internet applications that are ruled by Terms of Service and invisible to Google’s crawlers. Say what you will about the notion that the open Web would subsume all computing functions (surely that postdates 1995? It is, after all, at the core of Google’s current business model), but at least it was predicated on openness and interoperation. Today’s post-Web applications and services are built around artificial scarcity and raising the barriers of entry to competition.

In short, I’m surprised you’re giving in so easily. From a data, network and business perspective, today’s closing of the digital frontier makes perfect sense (indeed, we’re steering that way ourselves with magazine iPad apps. It’s simply the better market). But from a social, innovative, and macroeconomic perspective, open is almost always better. As Michael Wolff points out in his companion piece, it’s no surprise that media moguls (old and new) are pushing to regain control. What’s a surprise to many is that they’re getting it.

If “the Web” meant the open Web, why aren’t you shedding more tears over its decline?