announced the availability of a new streaming real-time Application Programming Interface (API) for all submissions, Diggs and comments on the site. Modeled after the Twitter Streaming API, Digg elected to use Tornado, the real-time framework built by FriendFeed and open sourced by Facebook, and Redis to power the API.Digg just
Will developers go for it? Though Digg's currency appears to be dropping fast, real-time streaming data from millions (?) of social media users, concerning links to content from all around the Web, has got to hold some interest for programatic analysis, UI innovation and publishing industry analytics. The flow of data coming through the API seems a little anemic, though. We spoke with some of the Web's leading data developers who see today's announcement in very different ways.
The way the implementation is offered is at the very least very contemporary. "Like Twitter's streaming API, it's another good example of the real-time web," says John Musser, founder of API directory and news site ProgrammableWeb. "We're seeing more and more 'event-driven' designs appearing in open APIs these days."
(Disclosure: Musser's blog is now owned by Alcatel-Lucent, who is also a ReadWriteWeb sponsor.)
Pete Warden, independent social graph analysis consultant, is optimistic for what this means to the ecosystem of mashup-friendly startups.
"Despite its recent troubles, Digg is still a massively popular tool for uncovering new stories. Offering a streaming API opens up all sorts of possibilities for mining that data, everything from spotting up-and-coming stories to offering detailed real-time analytics to content publishers. I know that startups like OneTrueFan are already salivating over what they'll be able offer."
Likewise, Tweetmeme founder Nick Halstead said today that he hopes to incorporate the new Digg streaming API into his much-anticipated new startup DataSift. "Our curation abilities on top of Digg submissions would be very powerful," he told us.
Is Digg's community exciting enough for it to still matter, though? Jeremie Miller, inventor of the Jabber/XMPP real-time/instant messaging protocol, is not so sure.
"Honestly, my gut instinct was 'digg is struggling to be relevant.' They should have published approximate flow rates with this too; some sparkly graph eyecandy woulda been nice diggs-per-second, etc. The stream is showing a flow rate of a few per second at best - Twitter's firehose is 500-1000 per second.
"Technically, their implementation definitely trumps Twitter's... whose firehose is about as un-sexy as it gets. Digg did a better job at that for sure - but it was such a low bar already.
"The only firehose I'd be surprised by and care about at this point would be Facebook's - but I expect we'll have first contact before that happens."
What Does it Mean for Digg?
If a wide range of third party startups and innovators can make meaningful use of Digg's streaming API, that could help form a virtuous circle that makes Digg more interesting to people again. If nothing substantial comes of the API, that doesn't bode well.
Witness, for example, the case of OneRiot. OneRiot is a well-funded real-time search service that put a whole lot of eggs in the basket of its ambitious API earlier this year. It's unclear that was effective in making OneRiot really stand out, though - and now the technically admirable, world-changing startup has put its tail between its legs this month and turned entirely into an advertising network.
Something like that may very well happen to Digg. Or, a streaming API could help make Digg hip and fun again.
What do you think, readers?