³ÉÈËÂÛ̳

« Previous | Main | Next »

Archrs - an everyday story of web development

Post categories: ,Ìý

Tristan Ferne | 11:56 UK time, Wednesday, 27 August 2008

The Archers

The Archers is a radio soap opera - "an everyday story of country folk" - that has been running for more than 15,000 episodes since May 1950. What I find particularly intriguing about The Archers in terms of the future of radio is that it is, in a way, happening in real-time. Every evening on Radio 4 the programme contains events that happened (approximately) "today" in - so if it's Easter in real-life then it's Easter in The Archers. It's also very topical - featuring real-life events such as outbreaks of foot-and-mouth disease or the World Cup. As a drama, each episode is split into distinct scenes - typically indicated by a fading to silence - a convention for audio drama because of its need for clear signposting. The other interesting aspect is the many distinct facets of each scene - which characters are involved, the time of day, the current location, the storylines that intersect at that moment and even the weather.

So earlier this year the Radio Labs team started to design and build a prototype website for The Archers based around the drama's scenes and facets. To start with we created a data model that represented what we thought was important in The Archers - including scenes, characters, character relationships (such as marriages or parent-child), locations, storylines and tags.

Data model for The Archers

Then we took the daily Archers podcast and built a web interface for a production team to use to segment and tag each episode. This allows a user to split the episodes into individual non-overlapping scenes and describe each scene with a title and synopsis and then tag it with the characters, location, storyline, weather and anything else they can think of. Once the user has finished this splitting and tagging, the individual scenes are extracted from the audio, saved as separate MP3 files (using if you're interested) and associated with the metadata.

Episode editor screenshot

Now we've got lots of scenes from the drama as MP3 files, timelines, storylines, characters and geography (we've even got maps!). What kind of Archers website could we build round this?

Episodes, scenes, facets, aggregations and storylines

First we created pages for every one of these facets; a page for every episode, for every scene, for every character, every family, every storyline, every place and every tag. Each of these pages aggregates the scenes that were involved and, at the second order, the characters/storylines/places/tags/etc that were also involved in those scenes. All of these facets are clickable so you can browse from a character to a storyline to another character and so on. And on each page the relevant scene's audio is aggregated in a single player - allowing you to catch up fully on that storyline or character, just listen to individual scenes or skip past the characters you don't like.

That's about as far as we got - building the skeleton of an enhanced Archers website allowing you to navigate around and through all the facets of this drama. But from this data we could also create custom podcast feeds for each of these pages - just want the Archers podcast for scenes featuring Eddie Grundy? Or a cut-down omnibus podcast with just the most crucial scenes for each week? No problem. We could even do a podcast feed for just those scenes featuring cows! One problem with a real implementation of this concept is that we only have 7-day Listen Again rights for the audio. So only the episodes and scenes for the last week would feature any audio. But thanks to the annotations we can create in the production interface and the superbly detailed episode synopses on the Archers website we should still be able to build something that works as an audio-less, text-enhanced service.

So...going back to my initial thoughts about how the Archers is in some kind of pseudo-real-time. My favourite idea would be to deliver the scenes from the Archers to the audience *as they happen* - so at 11 in the morning, when over in Ambridge Kathy has gone over to Home Farm for coffee, your podcast application gets updated in real-time with that latest scene. Or maybe with some clever shenanigans we could phone you up and let you listen in to the conversations in The Bull. For the dedicated Archers fans only I suspect.

Finally, some other thoughts:

Social history - The Archers often includes topical events and current affairs and generally represents the era in which we now live. So having an accessible, findable, addressable archive on the web going back 50 years could be a valuable social history resource.

Going smaller - Are scenes and characters the smallest units we can chop this drama into? Practically I think they are, though one could imagine going towards a finer grain of detail with marked-up conversations, characters' movements between locations, internal motivations...

What we built is only an incomplete internal prototype for the moment, there's lots more to do and there are lots of issues about how we'd do this for real - particularly around how we would practically annotate all these episodes. Maybe it could be a ?

Interestingly, though possibly not surprisingly, the ³ÉÈËÂÛ̳ has looked at this area before - Celia Romaniuk presented a paper on back in 2003 proposing an RDF vocabulary for describing the world of soap operas.

Comments

  • Comment number 1.

    Cool idea!

    How much time and people do you think it would take to fill the models and move from prototype to the real thing?

  • Comment number 2.

    Thanks. Filling the "static" data (e.g. characters, locations etc.) is basically done, some of it by some clever code that scrapes the Archers website. Annotating an episode took me about 10-15 minutes - but I'm not an expert and I wasn't annotating it that thoroughly. So starting from now it's not too bad and I think it's feasible. Annotating the entire Archers archive? That's another matter. Actually making the prototype live wouldn't be particularly hard but there's a lot of work around making a decent, usable site from this data.

  • Comment number 3.

    I'm assuming the "Archrs" in the title should be said with an Ambridge accent? :-)

  • Comment number 4.

    Nah, it's Archers 2.0! It's to take on Flickr :P

  • Comment number 5.

    Well spotted.

  • Comment number 6.

    Tristan,
    This is awesome. Until now granularity only went as far down as an episode. What you are talking about is taking granularity down to a level of an individual scene.
    This will allow users to listen in different ways -
    - By character
    - By location
    - By event (e.g. the marriage)
    or by any other term contained in the metadata.
    Have you looked at a metadata scheme or are you relying on the synopsis etc to pick up key words or user generated tags only?

  • Comment number 7.

    Patrick,

    We used the schema shown to manually classify each scene - and we're presuming that work would be done by production staff to get the quality necessary, though maybe it could be done by trusted fans.

  • Comment number 8.

    This comment was removed because the moderators found it broke the house rules. Explain.

Ìý

³ÉÈËÂÛ̳ iD

³ÉÈËÂÛ̳ navigation

Copyright © 2015 ³ÉÈËÂÛ̳. The ³ÉÈËÂÛ̳ is not responsible for the content of external sites. Read more.

This page is best viewed in an up-to-date web browser with style sheets (CSS) enabled. While you will be able to view the content of this page in your current browser, you will not be able to get the full visual experience. Please consider upgrading your browser software or enabling style sheets (CSS) if you are able to do so.