Last week's Eclipse DemoCamp in Kaiserslautern (
http://wiki.eclipse.org/Eclipse_DemoCamps_November_2011/Kaiserslautern) highlighted an important aspect of Social Media APIs, especially by Twitter sometimes more relevant to Enterprise and Social-BI than the client facing ones:
Streaming API (
https://dev.twitter.com/docs/streaming-api)
Among other interesting topics like EclipseSource replicating an entire Twitter-like "Munsters" style application based on OSGi and Eclipse RAP, one of the local hosts of the event presented this:
Whow is a webplatform, which refers its data from a tweetranking. The
webapplication follows a certain number of twitterers (based on data
from an external application) and evaluate their tweets according to
relevance. This produces a ranking of the hottest news at the moment,
which also can be personalized. By a theme explorer the user can add
more interesting topics and travel back in time per timeline.
All this has been realized in the backend using SMILA.Their so far mostly academic work on Semantic Filtering of Social Media Streams using Eclipse SMILA looked very exciting. And includes mostly Java-based libraries even companies like Google were interested in before. They would be happy to standardize some of those mechanisms. Most of them already are part of Eclipse SMILA:
wiki.eclipse.org/SMILA/GlossaryWhether or not, some of these would be in scope of one Social JSR or rather something else, I'm happy to discuss. To some extent the mechanisms serve BI and BPEL, but their work is focused on Social Semantics and BI, so there's certainly a bit of both.
I'll invite the lead of this project to the group, soon, too.
Werner