BBC RAD labs blog

HTML 5 and timed media

2009-08-12T11:27:11Z

HTML 5 work

HTML 5 is the next version of HTML, the markup language used on the web. Not all the details of the HTML 5 standard have been agreed, but many of the proposed changes and new features have already been implemented in existing browsers.

As part of our work on the P2P-Next project, we built a simple HTML5 demo that works in current versions of Firefox, Safari and Chrome: a sample of RAD's R&D TV with subtitles and chapter navigation. This will not work in current versions of Internet Explorer, nor earlier versions of Firefox etc.

What we built

This prototype plays video and audio without plugins, and allows jumping to chapters and 'scrubbing' within the content. It uses simple JavaScript framework to enable web page elements to be changed via individual HTML or CSS 'events', and for loosely-coupled publish/subscribe control of page components such as carousels. In particular, our JavaScript enables synchronised changes to HTML and CSS relative to a 'time parent', such as an audio or video clip, or even clock time. In addition, our solution needs to work with live events, whereby pages would be propagated in real-time.

Flash and other Rich Internet Applications (RIAs) provide something like this already via timeline scripting, but RIAs are 'black boxes', using compilers and obfuscators to hide code and data: great if you want to protect intellectual property, whereas we needed to provide a mechanism whereby data and the code acting on it were open and accessible. HTML 5 and the jQuery JavaScript framework gave us the tools we needed without requiring extra plugins or proprietary software.

Below we give more detail of the coding work done so far.

Synchronised media

On a web page it is sometimes desirable to make something happen relative to the current time of a video or audio track.

For example, subtitles change as a video plays.

YouTube Annotations adds something more dynamic: speech bubbles, notes and linked areas that can be set to appear in specific locations on a video, at specific times. This enables quite complex interactivity and navigation - in YouTube's example, a video about World War I leads to other videos about individual subjects such as tanks and aircraft, with extra notes that pop up as appropriate.

There are several open technologies available, or in development, for adding timed text (subtitles), but not more complex page changes and interaction. Annodex and parts of the Axmedis project go further in developing ways to integrate timed markup for annotation and indexing with streaming media. Annodex uses the CMML markup language, which can be either incorporated with, or referenced from, a stream.

The SMIL presentation markup language provides rich features for timed interactive media, and has been in use for more than 10 years, but it still has some limitations:

SMIL can be played in several media players, but browser support is patchy and the future of SMIL browser support is unclear
SMIL uses a presentation paradigm, and each presentation needs to be complete in itself, whereas our 'engine' needs to be event-driven, able to cope with individual live page changes, with viewers either joining a presentation part way through 'broadcast' or viewing the entire presentation after it has been published
existing SMIL implementations are oriented to elapsed (clock) time and user input events, rather than synchronising events with a 'time parent' such as the current time of a video².

Having considered the alternatives, we decided that our 'engine' would be most flexible and accessible if we used JavaScript alongside functionality developed as part of the HTML 5 specification.

The JavaScript we wrote does a simple job:

build a list of page-change event objects, parsed from incoming JSON
activate or deactivate each event, given its start and end time, relative to the current time of the event's time parent, which in our demonstrations is a video).

The HTML 5 audio and video elements remove the need for player plugins, work like any other HTML element in terms of styling and positioning, and standardise the programming interface for playback control. Less well known is that these elements emit a timeupdate event (at a frequency adjusted to fit available processing and memory) which removes the need to poll a player for the current time position. This makes media scripting far more efficient, since there is no need to run a loop or use setTimeout. In tests run on several machines we found that timeupdate events are emitted regularly and frequently (particularly in Firefox), whereas polling a media player for current video time is unreliable.

The HTML 5 addCueRange method will provide native support for callbacks (or events) at the start and end of a 'cue range'. However the specification is still under discussion and does not appear close to implementation.

HTML 5 video in the field

HTML 5 media elements are now supported by current versions of Firefox, Safari and Chrome.

However, the implementation of the video element has been dominated by the need to standardise codec support, in the same way that browsers currently support the JPEG, PNG and GIF image formats.

Google demonstrated their commitment to both the MP4 (H.264/AAC) and Ogg (Theora/Vorbis) with the HTML 5 media element at the 2009 Google I/O conference. Firefox 3.5 and Safari both support the video element, though for different codecs. Dailymotion has created a demo site with several hundred thousand videos encoded using Ogg Theora.

The biggest and least predictable change may come from technologies such as Comet or HTML 5 Web Sockets. These enable data to be 'pushed' to browsers from servers rather than vice versa. Push makes sense, in that enables updates without polling, but it challenges the HTTP request/response model used on the web and raises a number of security and editorial questions. In terms of our application, events could be pushed to the client browser as they became available. For example, a web page showing a broadcast from a live event could include live updated information about the festival or the band and song currently playing.

We've done this work in conjunction with the P2P-Next project. We're using HTML5 in the project to sync video and display extra information about the content - it's early days for us on this and there are a number of serious challenges before this becomes anything near mainstream - if ever. We hope it's a useful demo and look forward to feedback.

Quick install/build guide for RadioDNS

2009-07-19T11:17:06Z

Here's a quick list of how to get the RadioDNS application we released recently to build and run (on Ubuntu 9.04. Success reports for other distros or OSs welcomed.)

sudo apt-get install python-socksipy python-wxgtk2.8 python-dnspython python-setuptools python-dev

sudo easy_install objectlistview

wget -v http://github.com/BBCRadLabs/RadioVisDemo/zipball/master

unzip BBCRad*

cd BBCRad*

python radiovis_demo.py

And that should be it. Thanks for the useful posts on the radiodns.org list which summarised these steps.

RadioDNS Demo Application Released

2009-07-10T12:30:00Z

Following my previous post, Experiments with RadioDNS, I'm pleased to announce that the RadioVIS demo application is now available as open source software, under the Apache 2.0 license.

The application demonstrates radio station DNS lookup, querying for available services using DNS SRV records, and the RadioVIS protocol for displaying images and text messages. The code is written in Python and uses the wxWidgets user interface library.

Please note that this software is intended for those who want to find out more about the technical details of RadioDNS and RadioVIS, and not for end users. Also note that at the present time there is no RadioVIS service available for BBC radio stations. We are making this code available as part of our explorations with RadioDNS, and look forward to technical feedback.

To download, please visit our RadioDNS project page here. We'll put any updates to the code on the project page and announce them on this blog. The source code is also available on GitHub here.

For more information about RadioDNS, or to join the RadioDNS developer discussion group, please visit http://radiodns.org.

Episode 2 of R&DTV available

2009-06-03T11:52:37Z

We're really pleased that the latest version of our R&DTV show is now available to view, rip, remix, share, etc.

It's an experiment in creating an interesting TV show themed around technology, released under a Creative Commons licence, explicitly made available for you as viewers and end users. You can just watch and enjoy the show or do much more with the content using the Asset Bundle.

We actively solicited feedback after our first episode - and have taken a lot on board. We now have a richer experience, music, titles, etc. We have kept our approach deliberately low cost and it's still somewhat pro-am rather than TX quality. Baby steps as we learn what's the best way to create this content to allow reuse, but we think it's still an interesting show to watch in its own right.

Interviews in this episode include members of BBC R&D and RAD at the Maker Faire, David Kirby explaining the Ingex project, Matt Biddulph from Dopplr.com, and Jason Calacanis from Mahalo.com.

Get the shows here and let us know your thoughts here (rdtv@bbc.co.uk).

RAD's "How I did Agile" Day

2009-05-14T15:05:00Z

As this is my first post on the RAD Blog I should introduce myself. My name is Glen Ford, and I am the Lead Engineer/Technical Architect of the RAD Team.

RAD is a relatively new team in a challenging environment and it has always been Agile in the sense that constant change is a part of life in our environment and we all embrace that. However we have been searching to find the right framework in which to track and report without incurring too many overheads in time for team members.

RAD has to contend with the following:

Fluctuating workforce
Skill set specialisations
Team members not always on-site
Multiple projects with a mix of complexities and uncertainties
As with all teams, we're resource constrained - this means we need to multiskill and work flexibly

In order to address this we've undertaken some research into various Agile methods and frameworks. In the spirit of being Agile (uncovering better ways of developing software by doing it and helping others do it) we decided that it was a good idea to try and share that knowledge, and so the "How I did Agile" Day was born.

We managed to find speakers from both inside and outside the BBC who would be willing to talk on a variety of Agile topics. In total 32 people from a variety of departments and roles across the BBC attended a day of talks at Broadcasting House in London.

"Being Agile vs Doing Agile" - Rob Chatley

Rob is a Principal Engineer at Metabroadcast and one of the organisers of the eXtreme Tuesday Club. He also chaired XPDay, the UK's premier agile development conference, in 2007 and 2008.

Rob spoke about the core principles of Agile, and why we need to think about what we are doing. He warned of the dangers of good Agile techniques becoming lost in a sea of buzzwords. In a relaxed and well thought out presentation he covered the cost of development and how Agile can reduce those costs by delivery value early in the process.

In an interesting offshoot Rob also covered stress on the development team and how that changes between traditional and agile projects. Stress can be evened out through the approach of "failing fast", the idea that you tackle early in the process those tasks you know are potentially going to fail, so you have more time and a better chance of rectifying them.

"Kanban in Software Engineering" - David Joyce

David is an agile development manager with 20 years software development experience. Most recently David has been working for BBC Worldwide coaching teams on agile frameworks and practices including Scrum, Lean and Kanban.

David gave a very interesting presentation about applying Kanban to Software Engineering, from the origin of the term Kanban (Signal Card) to the principles behind it. David covered amongst other things the idea of pulling work through the pipeline, rather than trying to push it, combined with a way of managing capacity.

There was a lot of information in this presentation, and David has been kind enough to make it publicly available here.

"Agile in Distributed Teams" - Mark Coster

Mark Coster is a Delivery Manager with ThoughtWorks. He has worked with successful Agile teams on small single location projects to large 100+ teams distributed over multiple cities and continents.

Mark spoke of the challenges of running Agile in a distributed environment. He stressed the importance of interactions between team members and the customer. Mark also elaborated on the pitfalls that you need to be aware of, and some of the techniques you can use to mitigate them.

Techniques included avoiding silo creation by distributing roles across locations, the efficient division of labour and the creation of redundant roles to ensure better communication across sites.

"Behaviour Driven Development" - Adrian Lewis

Adrian is an Agile automated acceptance and performance tester with over 10 years experience with a penchant for dynamic languages and open-source tools, using Ruby/JRuby, Watir, Celerity, Rspec/Cucumber and JMeter.

Adrian gave a good talk on a challenging topic, introducing himself as an Extreme-Tester he described the change in approach to work from the outside in, by testing behaviour of an application. He also covered the challenges of defining requirements with a fluid imperfect language and how requirements are always a near miss, and the use of metaphors and stories to compensate. He touched on executable specifications using RSpec and conversations using Cucumber.

The talk sparked some good discussions on this approach and ways to engage it. And how this approach reduces the barriers to generating meaningful tests.

"Blue Skies" - Duncan Pierce

Duncan Pierce of Amarinda Consulting has been providing agile training, mentoring and leadership for development and management teams in the UK and further afield since 2001. Duncan is one of the founders of the XPDay conferences and regularly speaks at conferences and seminars in the UK and Europe.

Duncan spoke of the three types of projects - Blue Sky (experimental), Green Fields (engineering) and Brown Fields (legacy) - and how they each require a different approach. Duncan covered how the application of various techniques may vary in each case.

It was interesting to consider the project type and reflect on things such as: Risk, Success, Quality, Estimation, Testing and Pace. He also encouraged discussion amongst those present on what happens at the boundaries of these classifications and what you need to be aware of when working on "the edge".

The feedback on the day was very positive, and we were pleased that we were able to host such an event in the BBC and look forward to doing so again.

More photos can be found on the BBC RAD Labs Flickr Photostream

Many thanks to Chris Godbert and Katherine Kirk for their help in organising the day, and also to Sandi Ford for taking photos on the day.

Updates on R&DTV

2009-04-10T16:12:11Z

Quick post following up on our R&DTV experiment. So far we've had some useful feedback and fixed a couple of bloopers. Ooops.

The 5 minute Ogg file was a slightly old pre-final-edit version, where around 30 seconds of video were obscured. We've updated that now. It's in the same location on the FTP site - if you downloaded it previously, you might want to get it again. Or not, it's up to you.

All the other versions are fine. The versions up on the torrent sites seem to be the 30 mins version, so they're all good too.

The Guardian were very nice about these baby steps, we're on digg.com, TorrentFreak, currently on the front page of the Creative Commons site, and some of the lists seem to like what we're up to.

This'll be on archive.org soon and is on some of the torrent trackers I've seen.

Follow RAD or Backstage on Twitter for live-ish updates. Watching the search on Twitter for updates is fun for us as we see this spread.

Some of the commentators delve into questions that are most certainly not for us to answer - however, questions about codecs, next steps, etc, definitely are, so send them in to rdtv@bbc.co.uk and we'll try and answer in between eating chocolate bunnies.

R&DTV - New project from RAD and BBC Backstage/ R&D

2009-04-09T12:00:00Z

Today we launch a new trial aimed at exploring new ways to create, edit and distribute online video. It's called R&DTV. We've done this in collaboration with colleagues in BBC Backstage, part of BBC R&D.

It's a pilot show, designed to be sharable, remixable and redistributable. It's released under a Creative Commons Attribution (Non-Commercial) licence, and looks at interesting tech stories inside and outside the BBC. The first episode features Nicholas Negroponte (talking about the OLPC project), Kevin Rose from digg.com, Graham Thomas from BBC Research and Ant Miller / George Auckland from the BBC discussing the BBC Micro. Hemmy Cho and Rain Ashford were the producers.

]]> FTP site -'R&DTV is a monthly technology programme made up of interviews from knowledgeable BBC developers, project experts and experts from around the world.

The programming comes in 3 forms.

1. A brief 5 minute video, containing all the very best bits
2. A longer 30 minute video, containing deeper conversations
3. The Asset Bundle, containing everything we used and didn't use to make the videos above'

Releasing the assets as well as the 5" and 30" versions is something that's new for us. We think this is an interesting and possibly important experiement in creating video and audio specifically to be shared and remixed, from a professional content provider's viewpoint.

We want this to help shape how the BBC makes and encodes content - but this is early days. We've learnt a lot during the making of it ( eg don't include background music, even in passing, if you want to grant rights for reuse, etc) and are interested to hear thoughts about what works and doesn't work.

We plan to release another of these next month (in May) and maybe some more, depending on how this goes. We've deliberately tried to take a semi-pro approach to some of this. Doing it in a BBC edit suite with 'proper' TV cameras/ facilities would have made it much easier for us to make, but more expensive, and less interesting for other people. We'll be collating our thoughts and your feedback here and on the Backstage blog.

The 5 minute Quicktime version is here, the Flash file lives here, and here's the 5 minute Ogg version.
It's up on YouTube here and Blip here

More formats are available. See the file list for R&DTV, select one of the R&DTV asset bundles to remix, or look at the R&DTV readme

Please watch, share, remix, and feedback - rdtv@bbc.co.uk

URIplay released as free/open source software

2009-04-07T09:27:03Z

Hi all

RAD's really pleased to announce that we've released some recent work as free/open source software. It's a project called URIplay. The code was produced in collaboration with the BBC by an indie startup called Meta Broadcast who pitched the idea to us some time ago. We've been working on this together over the last few months and are delighted that it can now be shared. The reference code release bundle is available at http://ftp.kw.bbc.co.uk/backstage/rad/uriplay/uriplay-bundle-revision-2877.tar.gz. It's important to note that since our reference release, there's been a lot of progress on new features (mentioned below) and development of that is taking place over at http://uriplay.org/download/ - so go there for the full-fat code experience.

]]> You can investigate RAD's reference URIplay implementation, see the richer version operated by Meta Broadcast or read the API and find out more about the ontology

URIplay allows developers to access information on media brands or programmes (from a wide range of sources) using standard data formats and API calls. This means, for example that if a blog post refers to a piece of content on the BBC Radio site, and the URIplay API is used, people looking at the page from outside the UK, or outside the current rights window for that piece of content, could be provided alternative options - be that a pay-for aggregator, or information about when it's on next, or many more options (if they exist - This system can't create the rights to distribute content where the deals haven't been done!), and depends on upstream providers to make their content available in sensible formats and publish their URIs.

It currently works with a number of different content sources:

Podcasts and YouTube data through the same interface
Read data using standard RSS or RDF libraries.
Get context from Wikipedia,
Look items up via their iMDB links
Include data from the live web, via Twitter search. For example, what's hot on YouTube
Subscribe to links in iTunes (e.g., this aggregate podcast made from an OPML file of interesting stuff) and in Miro (e.g., this list of the latest YouTube videos discussed on Twitter--paste it into Miro's box titled 'Add Channel').
Follow links to other sources of data, such as MOAT and Freebase.

It's built on the fantastic work done by the team in BBC Audio and Music around /Programmes and is meant to be extendable for new uses we haven't thought of yet.

It returns RDF or RSS XML, and it's a strong follower of Web inventor Tim Berners-Lee's suggestions around Cool URLs not changing and Linked Data

It can operate as a federated service, allowing a number of different URIplay instances to run on different servers - querying the correct server for the correct information set. It doesn't store any data itself but queries its upstream data sources.

It's written in Java and we've had it running on Debian GNU/Linux and tested/ deployed on RHEL/Centos, but should run on any sane Java stack. A more detailed README around deployment/ real world war stories is coming this week.

We've used the Apache 2.0 licence, to allow the most widespread use possible for the code that we are contributing. This means you are free to use, modify or adapt this code for your purposes.

We're using URIplay in a number of our prototypes including our work on Totem/ Ubuntu, P2P-Next, and our recommendations system. We'll be going into detail about these soon.

One in Ten day

2009-02-23T15:28:18Z

In RAD we don't just work on interesting stuff ourselves, we also help others within the BBC to as well. RAD, along with the Research and Development's Innovation Culture team, have been working to roll out One in Ten across a number of areas within the BBC.

What is One in Ten? One in Ten is a way of gathering ideas and prototype proposals from colleagues - you may have read some my posts regarding this previously (eg Innovation and the red button )

It allows BBC staff to explore and investigate ideas that otherwise may have not been looked at as part of the usual workstream. The ideas explored may be things that are beneficial to BBC audiences or to the BBC's processes and systems. Recently our colleagues from the TV platforms team held a One in Ten day for BBC staff to hack and develop ideas. The day was themed "Games, consoles, quizzes and TV platforms" - Susan Richards from that team has written an update here

Ruhel Ali is a development producer working in the RAD and Innovation Culture teams.

Experiments with RadioDNS

2009-02-16T09:09:09Z

Hi. I'm a software engineer in RAD and I've been doing some work on RadioDNS which I'd like to share with you. As James Cridland and Nick Piggott explain in their article, RadioDNS - making your radio more intelligent, RadioDNS represents a way for broadcasters to provide an enhanced radio listening experience through a new set of internet services, which can be used by any AM, VHF/FM, or DAB radio that also has an internet connection or with internet streaming radio. These services include RadioVIS, which allows display of an image slideshow and text messages, RadioEPG, which provides an electronic programme guide, and RadioTAG, which allows the listener to tag or bookmark radio content, or give feedback directly to the radio station. The RadioDNS specifications, currently under development, are open for all device manufacturers and broadcasters to use, with a view towards future standardisation.

Radio Station DNS Lookup

A radio station broadcasting on FM includes digital data in the form of RDS that allows the radio receiver to identify the station and display the station name. As an example, the following example shows the station information broadcast for BBC Radio 1. The PI and ECC codes are obtained from the RDS data.

Extended Country Code (ECC)	ce1
Programme Identification (PI)	c201
Frequency	98.80 MHz

For BBC Radio 1, broadcast on DAB Digital Radio, the station identifier information is:

Extended Country Code (ECC)	ce1
Ensemble Identifier (EId)	ce15
Service Identifier (SId)	c221
Service Component Identifier within the Service (SCIdS)	0

RadioDNS describes how these parameters can be used to construct a fully-qualified domain name (FQDN) that identifies the station, in this case 09880.c201.ce1.fm.radiodns.org for FM, or 0.c221.ce15.ce1.dab.radiodns.org for DAB. A DNS lookup returns the canonical name (CNAME) for this FQDN:

nslookup -type=CNAME 09880.c201.ce1.fm.radiodns.org

Non-authoritative answer:
09880.c201.ce1.fm.radiodns.org  canonical name = bbc.co.uk

Availability of services is advertised through the use of DNS SRV records, as shown in the following table:

Service	SRV record
RadioVIS (Visualisation)	`_radiovis._tcp`
RadioEPG (Electronic Programme Guide)	`_radioepg._tcp`
RadioTAG (Tagging)	`_radiotag._tcp`

Capital FM have a RadioVIS visualisation service available for use with their iPhone client. To access this, we construct the appropriate FQDN from the RDS data and broadcast frequency: 09580.c586.ce1.fm.radiodns.org. The CNAME is found by DNS lookup to be vis.media-ice.musicradio.com, and a service record query against this hostname shows that the RadioVIS service is available at vis.musicradio.com, port 80.

nslookup -type=SRV _radiovis._tcp.vis.media-ice.musicradio.com

Non-authoritative answer:
_radiovis._tcp.vis.media-ice.musicradio.com     SRV service location:
          priority       = 0
          weight         = 100
          port           = 80
          svr hostname   = vis.musicradio.com

vis.musicradio.com      internet address = 194.70.58.122

Visualisation using RadioVIS

The RadioVIS protocol is an extension of the Streaming Text Oriented Messaging Protocol (Stomp), which can be served by a message broker such as Apache ActiveMQ. In order to demonstrate RadioVIS within the BBC I set up a local RadioVIS server for Radio 1 and developed a client application to display the content.

The server uses ActiveMQ running on Ubuntu, with content from the DAB LiveText feed for Radio 1 and the trial DAB slideshow service. The server provides two topics that clients can subscribe to, one for text messages and one for image URLs, which allows the client to be able to choose which notifications to receive. The topic names are constructed from the same station identification broadcast parameters we already used to construct the station FQDN, so for Radio 1 FM these would be:

Text messages	/topic/fm/ce1/c201/98.80/text
Images	/topic/fm/ce1/c201/98.80/image

After establishing a connection, the client sends a SUBSCRIBE Stomp message to subscribe to a topic.

SUBSCRIBE
destination: /topic/fm/ce1/c201/98.80/text
ack: auto

^@

Once subscribed, the client then periodically receives text messages and image URLs from the server, in the form of Stomp MESSAGE frames, where the body of the message body contains the specific message.

Text Messages

The TEXT message is used to display text on the radio receiver, such as programme and presenter names, or currently playing song and artist. Messages can be up to 128 characters in length.

TEXT

Example:

MESSAGE
destination: /topic/fm/ce1/c201/98.80/text
message-id: ID:radiovis-50431-1232528810717-4:6:-1:1:1741

TEXT On air now: The Chris Moyles Show^@

Slideshow Images

The SHOW message is used to display slideshow images. The message contains the URL of the image, which the client can download via HTTP, and optional hyperlink URL to be associated with the image. If given, the trigger_time indicates when the image should be displayed, or NOW to display immediately.

SHOW [,] []

Example:

MESSAGE
destination: /topic/fm/ce1/c201/98.80/image
message-id: ID:radiovis-50431-1232528810717-4:6:-1:1:1742

SHOW http://localhost/slide.jpg,https://bbcstreaming.pages.dev/radio1/ NOW^@

The RadioVIS demo application was developed in Python, using the wxWidgets GUI library and Jason R. Briggs's stomp.py module. We are planning to release this application as open source software soon.

For more information about RadioDNS, or to join the RadioDNS developer discussion group, please visit http://radiodns.org.

Thanks to Sean O'Halpin and Giacomo Schimmings from BBC Audio and Music, and Nick Piggott, Andy Buckingham, Adam Fox and Ben Poor at Global Radio for their help during this project.

Chris Needham is a Software Engineer in the RAD team.

Welcome to the RAD Labs blog

2009-02-13T17:38:34Z

Hi everyone. You're reading the first post on the new blog from the people working in RAD. We're a new, small team within BBC Future Media and Technology. I'm George Wright and I head up RAD. I have, and will continue to, blog on the main BBC Internet blog ( available here )

RAD works on near term prototypes, products and services across all digital platforms. We're part of a wider group encompassing Mobile, Audio and Music, and are based in London, W1. Our colleagues in Audio and Music already have an excellent blog at Radio Labs, as does the Journalism team, and we hope to emulate their approach to sharing new ideas and gain insight into your thoughts using the Web.

RAD completed recruitment and opened its doors in November 2008. We have a remit which stretches across all digital platforms and content areas, trying to help shape the BBC's thinking, encourage liaision with longer term research (both internally, within the BBC's fantastic R&D department, and externally, be it with academic partners, mainstream code releases or brand new startups), and release code and output which delivers new ways of tackling problems like distribution for AV content, service discovery, UI and UX for new platforms, and much more.

Whilst RAD was in gestation, we worked on new products and services, some of which have already been mentioned on the BBC's internet blog - our P2P-Next work, or the Ubuntu/ Totem collaboration. We'll be going into more detail about these releases on this blog.

We use agile methods to develop software. We've made some tweaks to the more formal ways of delivering using Scrum - lots of the people and companies we work with are external to the BBC, and so traditional physical standups won't work. We hope to share some of this thinking with you on here.

Our products are categorised into 3 phases - Alpha, Beta and Release Candidate. Very little of what we produce is designed for end users as a finished product, so we hope our categorisation lets you know what to expect. Alpha is the least fully formed - may barely run without crashing, but is functional and will provide learning. We'll explain these categories in more depth in a later blog.

We'll be posting fairly frequently to this blog, with a combination of code drops, new thinking, and general thoughts about our work.

George Wright runs the RAD (Rapid Application Development) team in BBC Future Media and Technology.