* Can cache for 24 hours, then must re-retrieve, if Guardian takes down article you don't want to still have it on the site * NYT - no corrections policy for data like there is for news * NPR - log crunching plus registration info (for api users) * Everyone has rate limits and attribution requirements * Approach violators in a nice way - we love what you're doing but we're a little concerned about X, so how can we work together * NPR - Intention is not to shut down people, do have TOS * Loosening control over the content a bit, but it's not giving away free reign * Guardian doesn't dog food their API yet b/c they've been building it * Strong alignment between UK geek community and Guardian readership * Commissioned developers they knew to write client APIs in Java, Ruby, etc. * Simon - Engaged with dev community early on, had internal hack day and invited external developers in (a FriendDA instead of an NDA) * Create's a lighter development process - site changes outside development group * NPR ate their own dog food first, to make sure it tasted good * APIs are indicitative of a good product * Stenger "Process and product are co-related and co-dependent" * At some point the taxonimies will have to collapse into a "lingua franca" * Daylife - we're at phase one of what will be a long evolution * NPR - meatdata standards problems for today, plus time since 1970 * Tags are functional - one tag = one page. * Guardian - tags applied when article is loaded in manaully by staff * Has disambiguation algo to tell people (Paris Hilton) from places (Paris, France) * Most of the content created by a dedicated Web editorial staff and put into a "Create Once Publish Everwhere" CMS * NPR - Most of the content they deal with is audio * Whitelist of contributors to the Guardian who they have license agreements with * Simon - content API is from their Web site (manual updates from print overnight), Web meta-data is a manual process for them * Split articles back into data - easier said than done * Organized Computation + Journalism conference * Use their API to make their own "edit of the Web" based on site whitelists * Let's publishers sidestep their own CMS * - Guardian tags driven by editorial decisions (so no tag for every MP in UK), site pulls up article and looks for tags on delcious and freebase (?) * Building a community of data wonks - hoping to get leads on better sources of data * Data Blog: Publishing raw data online via Google spreadsheets (easy interface for journalists, free Google APIs to the data) * For free you can use headlines, if you want text you have to join their ad network * Allowing commercial usage of content through API * Every piece of content tagged on Guardian's Web site * API based on full text search, can retrieve full text of articles * Simon - API launched on Tuesday of last week * It's a time savings for them, lets folks use the API while devs work on improving the infrastructure * API drives the NPR Web site, but extend the API to other developers in the building * Story based API - all content they have the rights to distribute * Retrofitted rest of site, launched to public in July * Jacobson - in charge all digital media development (anything but radio) * Went from a blog to a developer network * Interactive newsroom tech group - newsdrive applications # Get Me Rewrite! Developing APIs and the Changing Face of News
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |