BBC R&D

Posted by Libby Miller on , last updated

These are weekly notes from the Internet Research & Future Services team in BBC R&D where we share what we do. We work in the open, using technology and design to make new things on the internet. You can follow us on Twitter at @bbcirfs.

This week illustrates the lifecycle of projects in IRFS quite nicely, from inception to release and beyond.

Olivier and Denise each presented proposals for new work. Olivier's was "better, faster, smarter archiving of the BBC's websites. We agreed to start with a review of the state of the art in the domain, to get a better sense of who such archives are really for, and how they could be made better."

Denise's presented a proposal on evaluation of mood-based recommendations - taking the mood based similarity measures between programmes that the team have been working on, and evaluating them using iPlayer log data from Chris Newell. Penny, Sam and Theo have been helping plan this work.

Some projects are further on. The two-year TSB-funded project COMMA has just started and Matt Haynes, Chris Needham and Rob have been scoping it out and planning the work. Fi-Content 2 is undergoing a similar process, with Chris G, Theo, Penny and Joanne thinking about our planning meeting with Lancaster this coming week.

A more discrete piece of work also just starting is Snippets Waveform Display - a project to improve Snippets radio by visualising the waveform for the audio, enabling people to search the audio more quickly. Rob's been talking to potential new users of it, and thinking about the best way to navigate audio in the browser, again with Matt and Chris Needham.

Further on in its lifecycle is VistaTV. I've been doing a little QA on the marvellous dashboard, and getting to know the code, while AndrewN, Dan, Chris Needham and Thomas are working on getting it open sourced for the benefit of our project partners and anyone else who wants to use it. This is one way to successfully end a project or part of a project from our point of view. We can't maintain prototypes forever, but open sourcing the code means there's potential for it to live on elsewhere, as well as providing a record of the work in a public place as evidence of the experience and expertise of those who worked on it.

For all our projects there is infrastructure to be maintained, and Matt Haynes and James have been working through our new Puppet and Virtualisation setup, getting ever closer to having a fully automated system for deploying new virtual machines and software.

The World Service Archive prototype is going through another iteration. Tristan is thinking about measurement and what to focus on next. In the meantime lots of bug fixes and improvements are being worked on, and Chris L has been working on bringing together the flagging and exclusions code to make it easier for us to re-include content if we can get it cleared.

A subset of the Moods work has also been completed: Jana presented some results on facial emotion recognition, a collaboration with Smart Services CRC and the Queensland University of Technology. Running their software prototype over a subset of 5000 images from ELVIS, the BBC Information & Archives' internal digital picture library, gave a good selection of happy smiling people and another set of rather surly looking ones. Jana also prepared an even larger dataset for the EU AXES and the mediaEval benchmarking projects - instead of 1 week they now have 2 months' worth of BBC programmes plus metadata.

So rereading this, our projects remind me of a diagram from the mythology engine, with our projects as story arcs with smaller substories. Which makes the R&D website rather like a mythology engine connecting people, projects and events together in an easily navigable way. The R&D Website is now up and running, but not 'complete' - Olivier invited Nomensa to come and present the results of the web accessibility audit they performed on our R&D website. He says: "Our workshop with them was very interesting, and we were happy to learn that our site performed very well compared to the average website out there. Still, there are a handful of issues we'll need to address before we can consider our site's accessibility to match our standards." AndrewN adds "[the workshop] was a great refresher about the importance of building for all users, which might include thinking about things you can't see."

Finally, in 10% time, some of which may or may not form components of new projects in the future:

- I had another play with the Archers avoider and realised I need to learn how to solder (to fix my new LCD display to some headers for via the RasPi GPIO: otherwise looks straightforward - the Raspberry PI is great).

- Sean figured out how to export his Outlook calendar to emacs org-mode using a combination of Windows Scripting Host Jscript, his remote desktop drive attachment and Dropbox. "Heath Robinson, eat your heart out!"


Links, both from Tristan:

Your backup generator for the internet

Heroes of Science Action Figures! It's Benoit Mandelbrot!

Tristan says: "Which reminds me of a science joke. What does the B stand for in Benoit B. Mandelbrot?"

(I had to have it explained to me).