Yesterday’s Times Open event was a huge success. For a developer program in its infancy, turnout was excellent—140 attendees out of about 400 total applicants, according to Derek Gottfrid, the MC and principal organizer. Topics ranged from APIs to widgets to business models, but the message was clear: The New York Times is, suddenly, a platform company.
Keynote: Tim O’Reilly
For me, someone who read his first O’Reilly book (Programming Perl) 12 or so years ago, it was a thrill to finally see a Tim O’Reilly keynote (you can find my attempt at a liveblog in the previous post). In essence, O’Reilly threw down a challenge for The Times and every newspaper: to reinvent itself in the age of Web 2.0, and to discover a new business model while doing so.
The talk summarized the basic tenets of success on the web, with particular attention paid to the concerns of a newspaper. O’Reilly criticized the Times bluntly for current shortcomings, and suggested ways they could improve. Still, not all problems have easy answers. The biggest challenge for The Times is to balance the openness that characterizes success on the web with the propriety and tradition of a newspaper—and not just any newspaper, but a centuries old one: The Old Gray Lady herself.
I’m not going to cover this talk in detail because I think the liveblog below does a pretty good job. Suffice to say that the gauntlet has been cast.
At this point each developer (or development team) ran through their API (or APIs). It was a brief run through, and most of the information is available on The Times’ developer site, but I highlight the parts that interested me the most.
To be released next week, this will be The New York Times firehose. It will provide a real time feed to track content as it is published. At first this will feature a trimmed down data set (just articles to begin with, I think), but it will eventually include all online Times content (including blog posts, for example).
The feed will be cached for up to a second, and they are happy to have people poll it as often as they like (they kept saying that we should poll it every second, but this is flawed for a couple of reasons: 1. it will only feature about 300 items a day, so polling it every second it serious overkill; 2. we are limited to 5,000 requests per day per API key, so polling every second would put us over the limit in a little over an hour).
Article Search API
I was most intersted in Derek’s future plans for the API:
- Open Calais integration (which is a great idea and a thought I had myself)
- Related document search
- More fields
- More article data
- OpenSearch RSS/JSON
After the API reviews we broke for lunch. I got to meet Steve Myers from The Poynter Institute, Gilad Lotan of Global Voices Online, and a handful of others (if you were one of the others, please leave a comment and identify yourself!).
The next session was an overview of R&D at The Times, including a presentation of shifd, which was incubated there. I actually missed this presentation because I got to go on an unofficial (clandestine?) tour of the R&D departnment, which I would happily take over a presentation any day. I’m not sure how much I’m supposed to talk about this, but they are working on some very cool stuff, including a fully wired living room that uses your cell phone to identify you and show you localized content tailored to your preferences. They also have a killer view from the 28th floor of The Times building.
Jacob Harris’s presentation was one of the most intruiguing of the day (and one of the most difficult to put together, I imagine). He started with a premise that interactive news can be described as a triangle whose points are:
- Data is the raw material - that from which news is forged
- Story story gives data structure; in traditional media, this is what turns data into news
- Interaction gives news dimension, makes it compelling, and gives users a reason to use it
Often, interaction has thus far come down to navigation (navigating a data set, for example). Jacob thinks we can do more, and hopes to see some really interesting resources for Interactive News grow out of the developer community.
The Alcohol API
When the day was over, we gathered over food and what the organizers insisted on calling cocktails but what I call beer and wine. (Brian Kennish had termed this the Alcohol API.) I finally got a chance to meet Derek G at this point, which was nice because he had some nice things to say about my nytexplorer app, and I had some nice things to say about his API.
We ended up going to a bar after that, and things devolved and the group waned. In the end it was just me and David Billingham. Unfortunately I had to run to catch a bus, because I wished I had more time to spend with him.
The New York Times recognizes the need for a news platform, and wants to fill that void. Their developers understand that Web 2.0 is about openness, and are committed to opening up The Times data to outsiders (whether or not the suits—and they do wear ridiculously nice suits—agree remains to be seen). They want to create a community of hackers around their APIs, and invite community feedback to drive future innovation. The Times executives should be supremely grateful for having employees with such vision at their organization.
My most sincere thanks go out to Derek and the whole Times dev team for hosting this event. I am thrilled to have been there. I don’t care what people say—this event was worth every minute. (I was lucky to get in at all—I was literally the last person on the list, and only got on there because I sent nytexplorer to Derek on Thursday afternoon).
Until next year?