The next version of ODDT (0.9.5) has been submitted to the AppStore, and should be available within a week or so. The most noticeable addition is that the main screen how shows a ticker along the bottom, which cycles through recently endorsed factoids throughout the whole topic selection. Also the scrolling behaviour has changed: the topic panels now always scroll left-to-right regardless of device orientation. A new topic has been added – Global Genes – and there are some internal improvements to help ensure that user contributions via Twitter are picked up in a more timely manner.

The ticker has been implemented in response to the gradually increasing number of topics supported by ODDT: you may be interested in quite a few of them, but opening up each of the panels regularly to check the highlights is a little bit more tedious than it needs to be. So every few minutes, the app asks the server to furnish it with the 50 most recently endorsed (or auto-endorsed) factoids, from all the topics. A timer cycles through this list, and displays the text on the bottom of the screen, along with the graphic for the topic that it belongs in.

Tapping on the ticker display brings up a detail view of the factoid, e.g. a recent auto-endorsed tweet with a link to an article about green chemistry:

This is the same detail view that can be accessed from within the topic viewer, with the exception of the extra button along the bottom right: the circle button is a shortcut for jumping to the topic display, which saves you from having to hunt around for it on the main screen.

The factoids that are listed by the ticker are limited to those that have a positive endorsement, i.e. at least one person has sent out a tweet to endorse it, or it originated from a tweet that included the #ODDT tag (which triggers an “auto-endorse”). The list is sorted according to the time of most recent endorsement, which means that whenever you endorse a factoid in any topic, it will bubble up to the top of the list in everyone’s ticker display.

At a later date we will probably add a feature to specify which topics should be included, but for the moment the throughput is low enough to include them all.

Another important modification that has been rolled out into the current version regards the sometimes-flaky behaviour of the Twitter API. The ODDT server regularly polls Twitter, looking for certain hashtags, but for reasons unknown, certain users go through blackout phases, i.e. their tweets never make it into the index. It’s intermittant, so probably has something to do with the well publicised scalability problems that Twitter is known for. As a slightly hacky workaround for this, the server now takes advantage of the fact that many of the read-only requests from the client include the Twitter handle of the user. It makes a note of this, and polls the user specifically, looking for any relevant tweets that it may have otherwise missed. The hope is that this dual redundancy will reduce or eliminate problems with content not showing up.