The New York Times APIs: Shimmers of promise in early uses, but real work starts tomorrow
Ever since The New York Times began rolling out a series of APIs four months ago, I’ve been eagerly anticipating what might come of them. Tomorrow’s event at the newspaper’s midtown headquarters, Times Open, is intended to encourage development around the APIs and could point in a few new directions for news delivery and consumption.
If you can’t tell your MySQL from PHP, this is what you need to know about APIs. The acronym stands for application programming interface and refers to the way that applications — like Twitter and UPS package tracking — interact with each other. Those sites have a ton of data you might want to access, but perhaps not at Twitter.com or UPS.com. You might, instead, prefer to see your friends’ latest tweets on your Blackberry. Or you might want to view the location of your package on Google Maps. Those mashups, as they’re often called, are only possible thanks to APIs.
How might the Times’ APIs be used? Well, one of them provides access to all of the newspaper’s movie reviews. Wouldn’t it be helpful, when making you’re next Netflix selection, to see the latest recommendations by A.O. Scott without navigating to the Times? Netflix just released its own API, so this application now, blissfully, exists.
There may also be a business model here: Marc Frons, chief technology officer at the Times Co., has said they plan to charge for commercial uses of their APIs. (That link leads to a very helpful Poynter piece on this same topic.) So if Amazon wanted to add the Times best seller lists to its site, customers might benefit from the additional information, and the Times could get paid for its work.
The Times has released eight APIs for its content ranging from campaign contributions to Congressional roll call votes to every article published since 1981. The article search API is the most evidently promising, and it has already spawned some stunning visualizations by Jer Thorp, an artist in Vancouver. The one above tracks frequency of the words “hope” (in blue) and “crisis” (in gray) from 1981 to 2009, starting and ending at midnight. This one does the same for “sex” (beige) and “scandal” (pink):
And for each year between 1981 and 2009, Thorp has also charted the most frequently mentioned people and organizations in the Times and their relationships among each other. Like this one for 2008:
Here’s a bigger version of that image. Check out his Flickr set for plenty more beautiful visualizations of the Times oeuvre, and Thorp explains the technical details in a set of posts on his blog. (That’s where I get lost.)
There are a few other, early uses of the Times APIs, three of which have been highlighted here. A script by Times developer David Seguin leans on one of the APIs to offer a new way of viewing comments on the Times site. And this code makes it easier to use the Times’ roll call votes. But that, as far as I can tell, is pretty much it, which means the potential of these APIs hasn’t really even begun to be realized. I’ll be at Times Open tomorrow, hoping to find some of the more exciting possibilities.
The excitement over the growing number of API’s being made available by the New York Times (and others) is driven by the wide range of possible value-added applications.
But I think we should not abandon a healthy level of some skepticism.
We should keep in mind that these API’s are almost certainly not being offered (purely or even mainly) out of a sense of altruism or public service. The publishers will almost certainly charge for these services when and if the demand grows sufficiently – it makes no sense not to. (Keep in mind that onlline newspapers are getting increasingly desperate for funds, and you’d be naive to assume this would be an exception.)
Just look a the terms of service. Some TOS’s outright restrict the use to non-commercial purposes. Others, like the NYTimes, (appear to – for now, anyway) allow some commercial uses, but at the same time prohibit any user of the API from employing that API in anything that competes with the NYTimes (current *or future* offerings). If you realize that one of the main motives of those who release the API’s is to see what good ideas others can come up with, is it such a stretch to assume that these same companies will be eager to implement good ideas themselves, regardless of where they came from? And then notify you that you’re now competing, in violation of the TOS?
In any case, even without the incentives just discussed, I’d be very careful about building a business (at least one that requires more than a trivial investment) around services that are provided by a third party, especially when that party’s terms of service declare they can change anything at any time they want. That’s just good business sense.
Just my $0.02.
See also nytexplorer.com – just launched!
Thanks, Taylor. It looks great and seems like an improvement, in certain circumstances, over the Times’ own search function. I just gave it a tweet and hope to meet you tomorrow. —Zach