Saturday, April 17, 2010

Chat with with Daniel Jacobson

Earlier today, I met up with Daniel Jacobson, one of the architects of NPR's API and an all-around good person. We had been in contact via e-mail over the years, but to meet in person in Chinatown here in DC was a treat. And for all the micro slices of communication that we all do on Twitter and other social networking sites, it's easy to overlook the simple pleasure of getting together and talking shop over lunch. As an independent developer, someone who is not affiliated with a station, to meet over a sandwich and have a heart-to-heart with a colleague is even more of a treat. So, in case you haven't done this in a while, I encourage you to look at your list of friends on Twitter or Facebook and invite someone out to lunch.

If you are already plugged-in to the #pubmedia discussions, most of what we talked about was not so much new, but rather it was a friendly reinforcement of ideas and initiatives that have been mentioned before.

I was particularly pleased to have been reminded of how well the "NPR API Ingest" project is coming along. With this project, NPR and member stations are working together to actually take station-produced stories directly into NPR's content management system and syndicate this content back out through the NPR API. See: A recent tweet by Daniel Jacobson shows the "NPR API Ingest in action... - - This is a story from N3, ingested into the API, and pulled back out by WBUR." So a Northwest News Network story goes into NPR's web site, then comes out via the API and is displayed on WBUR's web site: with text, audio, images and all.

Here is the story on
and the story on,1025,1032,1035

This is a big thing!

One follow-up question I might have asked (as a possible post for the Inside NPR blog) is how does NPR API Ingest work? What are the required fields for a story? Are stories saved as an xml feed and then taken into NPR's archive automatically? Is there some editorial judgement or massaging of the story that happens before it becomes part of NPR's database? If someone were to write an app for a public radio station and wanted to anticipate the ability for NPR to ingest station content, what would they need to know?

Daniel talked about an upcoming blog post at the Inside NPR blog on the metrics and usage of the NPR API. It should be interesting to find out what unexpected places NPR stories are showing up (and eventually station stories) as a result of making the NPR API publicly available.

We talked about a thread on the Public Media Google Group started by WGBH's Chris Baer outlining his work with NPR content and Solr as well as people's suggestions for ways in which the NPR API could be used differently. Daniel agreed with me that "giving NPR Music some love" was an idea whose time has come. I expressed the thought that I may have clouded the subject with my brain dump on the idea of an NPR Music Megaphone app, and really should have pointed people to the Magnatune Radio: and said, "imagine this for NPR Music."

We also talked briefly about discussions happening at the CPB and around the system (likely similar to this post by the Vice President and Chief Content Officer at Louisville Public Media Todd Mundt on "The Challenge of Serving Audiences Where They Are") where organizations are asked to consider making digital media a priority commensurate with their on-air broadcast.

We discussed some pet projects like Daniel's NewsMap: and how it was good to simply work on a coding project for fun.

And there was more... four pages in my Moleskine, jotted down thoughtfully while on the Metro. And that was the result of just one hour over a vanilla malt and a grilled cheese sandwich at Potbelly's. Like I said, look through your contacts in your rolodex or social network address book and meet for lunch. It can be a really good thing.

1 comment:

Daniel Jacobson said...

You forgot to mention your pending move to Paris! That was definitely an interesting part of the conversation.

Regarding the ingest process, we do plan to post to Inside with how the system works. The high-level view, however, is that we authenticate the station, coach them on preparing NPRML or MediaRSS feeds (with some namespace extensions), and provide details on how they can post the content to our API. More details to follow soon.

Looking forward to the next lunch, before you head overseas...