To get a quick idea of how the Open Drug Discovery Teams (ODDT) project is coming along, check out this screencast. The ODDT app is at a fairly basic stage, but it is taking form nicely. The screencast shows the opening of a topic – Tuberculosis. The inside cover for each topic now has three sections:
Selecting the Incoming tab will show a stream of factoids that have been harvested from the twitter stream, on account of having one of the hash tags that is associated with that topic (in this case, #tuberculosis). This content updates in realtime, quite rapidly. Many (most?) of these incoming tweets have little relevance to people who are actively interested in relegating TB to the history books, but some of them are useful, typically on account of including links to relevant information, or even structured chemical data.
While this information is harvested by the back-end server, it is the role of users of ODDT to identify which of these factoids are good or bad. This is done by emitting tweets, from the mobile device, which the screencast demonstrates: emitting an endorse tweet will add a thumbs-up, while a disapprove tweet will add a thumbs-down. This is displayed on the page view. The endorse/disapprove tweets have the #ODDT hash tag, and some other identifiers so that the server can recognise them. You can check out some of these test tweets by looking at my account, @aclarkxyz, or the development testing account, @oddttest.
When a factoid is crowd-approved, i.e. it has more endorsements than disapprovals, it qualifies for the Recent selection. While the Incoming tab displays all of the tweet-derived factoids in order of most recent first, the Recent tab just displays those with a net positive endorsement total, which is essentially a first-pass curation result, since everything in there has been approved by at least one user. These can be voted off the island, so to speak.
The third tab, Contents, also shows only factoids that have a net positive endorsement, but they are ordered by most popular first. This presents a way for people interested in the topic to get up to speed on the open data that is available, without having to wade through too much noise.
The screencast is a demonstration of functioning software. While the app is not yet ready for the AppStore, and the server is running on a development box firewalled away from the greater internet, the feature set is getting close to a minimum viable product. So watch this space. Or even better, get in touch, and tell us what you think.