Of the Web vs. On the Web

Nov 15 2007

News organizations spend a lot of time talking about what they’re doing “on the Web,” but there’s another phrase that’s more important from a long-term perspective: “of the Web.” There’s a significant difference, although it can be abstract to folks who don’t spend a lot of time thinking about online. A simple difference between the two phrases can be illustrated by news organizations that simply put their content online and those that do go beyond that and use the capabilities of the Web through links and other resources. (Jacob has a related example for software).

Which brings me to CARMA. Today, the Post and other news organizations published stories about CARMA’s efforts to track carbon-emitting power plants worldwide. The Post published a small map with the story in the paper, while other news organizations alerted readers to the existence of the site. This is “on the Web” stuff - valuable, but not much more so than what’s in the paper.

A couple of days ago, in a time-honored tactic, CARMA sent out advance notice to journalists and offered a sneak peek at the database. Out of curiosity I visited the site, which is nicely done (although running a bit slow today, likely a result of the traffic). And then I saw the bit that the press release didn’t talk about: the API. In short, for those folks who want to use the data in some way that CARMA didn’t think of, there’s an option that delivers information over HTTP. The API, and the options it provides, is definitely “of the Web.”

APIs are something don’t come naturally to news organizations, WPNI included, but they should be a larger part of our thinking. Consuming data via API when available is a no-brainer; it replaces the time-consuming and often repetitive tasks of requesting and processing data. You change your data processes when the API changes. Producing APIs, on the other hand, is a slightly tricker business, since there are plenty of news folks who get nervous when you talk about providing published content to others with no inclination of how those people will use it.

In some cases we’re already doing APIs, of a sort at least. Take RSS feeds, which can be consumed and transformed for different purposes. But few of the database applications produced by news organizations have such feeds. The current widget rage is nice, but it hides the data from these widgets behind Flash and other wrappers. The appeal is control - we know exactly what people are doing with our widgets, because we control their actions. The loss is the uses we haven’t even thought about.

What really struck me about CARMA’s API is that the press release to news organization didn’t mention it at all. That’s because it isn’t intended for us, and that should worry news people. Think about it: the groups that have useful and interesting data are building alternate routes around the news media, maybe because they were told that APIs are the next big thing, or maybe because they know that getting their information in front of as many people as possible requires a certain amount of letting go. Either way, Web users are going to find CARMA’s data - some of them driven from a news Web site, but some of them without ever having visited one. What’s connecting those users and that data is HTTP - the “lingua franca of our age,” as Jacob puts it. Being “on the Web” - particularly for database applications - is not enough. We must strive to be “of the Web.”