![](https://ichef.bbci.co.uk/images/ic/400xn/p0fd5m8r.jpg)
Welcome to the regular update from the Internet Research & Future Services team in Βι¶ΉΤΌΕΔ R&D, making new things on, for and with the Internet. Βι¶ΉΤΌΕΔ R&D is on Twitter at @bbcrd
Atomised News
Whilst looking at longer terms objectives and planning for the Atomised News project, Barbara investigated the options for the large scale, public testing of our prototypes. Based on the current options and requirements, the team has made the decision to build the prototype for public testing as a web app, rather than as a native application.
Anthony was busy working on the integration of , which is structured data CMS and has been investigating for app development. He's also been studying the .
Andrew and Tom have been focusing on the development of our 4th iteration prototype - a semi-functional prototype with a new UI and interaction patterns that include changes based on the feedback from our most recent user testing of the 3rd iteration. The UI went through several iterations but we're very happy with the outcome. Andrew was preparing all the assets with content written by our journalist, Lauren, until he was struck down by flu. Tom has been developing the prototype as much as possible using and also had initial meetings with the Child of our Time production team to establish requirements for that project.
EBU DevCon 2015
Last week, Sean and Chris were at the where we hosted a workshop on . The workshop explored the architecture of CPA; the mobile, HbbTV and RadioTAG libraries and how CPA could be implemented. You can find the virtual machine image which we prepared for the workshop here: Thomas also gave a talk at the conference on our experience with using and container technology in building some of our content analysis prototypes.
Discovery
In the past few months the Discovery team has been experimenting with a graph-based, related-topic discovery prototype, which we have called Halo. We've also been looking at the discovery of similar articles within a corpus using vector-space models which combine binary and continuous attributes. This sprint the Discovery team has been comparing results of the two approaches and experimenting with how the two can be used together. We have also been experimenting with spike detection by monitoring a graph of semantic connections over time.
SUMMA
IRFS has been meeting with to see how we might contribute to a new EU collaborative project called which will build an automated media monitoring system. The system will be designed to monitor hundreds of international TV channels and radio stations, as well as online articles and social media, identifying trends and breaking news stories. Working in several languages the system may provide applications for our work on and media processing pipelines.
Goodbyes
This week we said goodbye to our two graduate engineers, Tom and Peter who have both been promoted and are moving on to their next placements in R&D.
What caught our eye