ubiquitous i a: building for change and web 3.0

BBC MMVIII

Ubiquitous Information Architecture: Building for change and Web 3.0

Chris Thorne

Information Architect

BBC

BBC MMVIII

Me

• An Information Architect.• Playing in metadata and the web since 1999.• Before that a Multimedia Research Scientist

(for Philips Electronics).

BBC MMVIII

BBC MMVIII

About the BBC

• The largest broadcasting corporation in the world• Mission: To inform, educate and entertain.• Public service broadcaster, established by a

Royal Charter and funded by the licence fee that is paid by UK households.

• Services include: 8 national TV channels plus regional programming, 10 national radio stations, 40 local radio stations and an extensive website, bbc.co.uk.

BBC MMVIII

My Part of the BBC

- Provides web technology for the whole of the BBC.

- Primary customers - BBC Audio, Music and Mobile.

BBC MMVIII

My Part of the BBC

/Programmes – a multiplatform dynamic semantic publishing service to support all programmes broadcast by the BBC.

BBC MMVIII

My Part of the BBC

/music – a multiplatform dynamic published semantic service to support music activities by the BBC.

BBC MMVIII

My Part of the BBC - Multiplatform

PC, Mac, Mobile (WAP, iphone et al) and future devices and platforms.

BBC MMVIII

As Information Architects

We play a key role in determining:

• How information systems interact.• How information is stored.• How users and businesses get the most out

of these systems and services.

BBC MMVIII

The world around us is changing……

‘Ever more pervasive, ever harder to perceive, computing has leapt off the desktop and insinuated itself into everyday life. Such ubiquitous information technology “everyware” – will appear in many different contexts and take a wide variety of forms, but it will effect almost every one of us, whether we’re aware of it or not”.

[‘Everyware: The dawning age of ubiquitous computing’ – Adam Greenfield]

BBC MMVIII

Mobile Devices

BBC MMVIII

Kiosks and multi user devices

BBC MMVIII

Public Information/Advertising

BBC MMVIII

Discound/Credit/Travel Cards and readers

BBC MMVIII

Radios using the web

BBC MMVIII

Even TV is changing – the web is on TV

BBC MMVIII

And how is the web?

:-)

BBC MMVIII

Where we are now – web 2.0

BBC MMVIII

Computers can’t help much

BBC MMVIII

The Apple=Apple=Apple problem.

A computer struggles to understand the different ‘Apples’.

BBC MMVIII

Web 2.0 - Too many Silos.

If you want to use a web 2.0 service in your website, you have to persuade your engineers and developers to learn a new API, with new data formats and new interfaces.

BBC MMVIII

The Semantic Web is not just academic….

BBC MMVIII

The challenges include everyday life.

• Its hard getting to work.• Its hard shopping for stuff.• Its hard to socialise……

These user journeys are part of our everyday lives and yet the web and current everyware isn’t that much help.

BBC MMVIII

Semantic Web – A web for machines.

• Trying to build a web of things (including documents) that machines can understand.

• So machines can help us solve everyday problems more easily.

BBC MMVIII

Idea 1

‘Everyware’ is here now and will increasingly influence our everyday lives.

BBC MMVIII

Idea 1 continued

But it needs a lot of improvement.

(we still have a job, yay!)

BBC MMVIII

Idea 2

The Semantic web will succeed.

It will be web 3.0, web 4.0 or web 5.0……

Semantic web will power significant amounts of everyware.

BBC MMVIII

Idea 3

Information Architects need to engage with semantic web and ‘everyware’…..

BBC MMVIII

Idea 4

We can build now for a semantic future, where the web is supporting a variety of information services and devices.

BBC MMVIII

Idea 4 continued

….even if we’re building microsites.

BBC MMVIII

Idea 4 continued

….even if they are statically

published.

BBC MMVIII

…..The BBC and others are doing this now.

BBC MMVIII

BBC MMVIII

Semantic Web – machine readable data

Trying to explain:• Who• What• Where• When

To machines……

BBC MMVIII

The BBC Semantic Ambition:

BBC MMVIII

Progress so far:

BBC MMVIII

Bbc.co.uk/music

• Powered by musicbrainz data.• 440825 artist profiles and growing.• BBC is not solely responsible for maintaining

and growing the database - but is able to contribute BBC effort to the benefit of the whole web.

BBC MMVIII

Semantic Web - How does it work?

• Standardised building blocks – XML, RDF, Ontologies, HTTP

• Allows easier exchange, crawling and use of data.

BBC MMVIII

Semantic Web - Layers of a cake?

BBC MMVIII

Statement 1: Chris - Knows - Michael.Statement 2: Chris - Lives in - LondonStatement 3: Chris - Age - 34.

Ie simple statements about things. Aka ‘triples’ - dbpedia.org has 116.7 million triples.

Semantic Web - Example RDF statements

BBC MMVIII

Semantic web - Linked Data Principles

1. Use URIs as names for things.

2. Use HTTP URIs so that people can look up those names.

3. When someone looks up a URI, provide useful RDF information.

4. Include RDF statements that link to other URIs so that they can discover related things.

Tim Berners-Lee 2007

http://www.w3.org/DesignIssues/LinkedData.html

BBC MMVIII

In other words

• Based on standard web etiquette. BUT…..• Without unique URIs and persistent URIs,

people aren’t going to like using your data.• Each time they want to know something its in

a different place? You’re having a laugh :-).

BBC MMVIII

Semantic Web, Restful design, Google juice

Publishing semantic data, like maximising google ‘pagerank’, relies on persistence.

• The URIs you use should be persistent.• Otherwise your google juice/other

companies/users/machines get lost.

BBC MMVIII

Google pagerank maximising – what we know...

• Do good stuff.• Practice RESTfull principles.

– Put your ‘good stuff’ in one place only.– Keep it there forever.

• Get your H1s right.• Tell others so they link to it.• Get companies with a high page rank to link

to it.

BBC MMVIII

Restful design principles in brief:

If the contents of a page changes state, the URI changes.

This means personalisation using cookies can be bad – if a page has changed state without changing the URI.

Use AJAX/Flash with caution.

[see http://en.wikipedia.org/wiki/Representational_State_Transfer]

BBC MMVIII

Semantic Web – machine readable data

Trying to explain:• Who• What• Where• When

To machines……

BBC MMVIII

Who and What

…..The problem is they change……but we want them to remain at the same URI…..

BBC MMVIII

Who and What change - some examples:

Prince, symbol, slave etc

BBC MMVIII


Victoria Beckham, Victoria Adams, Posh Spice etc

BBC MMVIII


• Eg. What can similarly change over time – PC started off as IBM PC, now there are Dell PCs, Acer PCs, the term PC is more known than IBM PC.

• How to futureproof and allow for changing names and changing products?

BBC MMVIII

Ambiguity is useful (sometimes)

• Rather than use a persons name to identify them (which may change) use an identifier.

• This identifier must be persistent.

BBC MMVIII

Identifiers example - GUIDs

• type of identifier used in software applications in order to provide a reference number which is unique in any context.

• Total number of unique keys (2128 or 3.4×1038) • 16-byte (128-bit) number.• most commonly written in text as a sequence of

hexadecimal digits such as:

3F2504E0-4F89-11D3-9A0C-0305E82C3301

BBC MMVIII

Identifiers to the max: Webscale identifiers

• The same identifier is used by several organisations to uniquely identify the same thing.

• Eg Music brainz identifier for Victoria Beckham is:

ce8dbfe4-3ce7-41e9-9bb3-13e35c053318• Eg Music Brainz identifier for Prince is:

070d193a-845c-479f-980e-bef15710653e

BBC MMVIII

http://musicbrainz.org/artist/070d193a-845c-479f-980e-bef15710653e

BBC MMVIII

http://www.last.fm/mbid/070d193a-845c-479f-980e-bef15710653e

BBC MMVIII

http://www.bbc.co.uk/music/artists/070d193a-845c-479f-980e-bef15710653e

BBC MMVIII

070d193a-845c-479f-980e-bef15710653e

• The musicbrainz identifier for the artist ‘Prince’.

• It doesn’t change, even if he changes his name. This allows carefully built URIs involving Prince to be persistent.

• It is webscale because it is understood by more than one website.

• This enables BBC, Musicbrainz and LastFM to exchange data about an artist easily.

BBC MMVIII

In fact, identifiers are useful for all websites……..

BBC MMVIII

Google and identifiers in URLs.

• Persistence is most important.• Other sites link to you = higher page rank.• Users can bookmark and share persistent content.• If you move your pages to new URLs you lose

google juice.

[ok you can put in a HTTP 301, but its still bad, as Tim BL says ‘Cool URIs don’t change’]

[see: http://www.w3.org/Provider/Style/URI]

BBC MMVIII

Identifiers – Helping us cope with change.

• Avoid losing your google juice when names change.

• Unique alpha numeric codes used to uniquely identify a thing.

• Must be persistent.• Use the same identifier across several

companies to make information exchange easier.

BBC MMVIII

How BBC designs Persistant URIs - tips

Persistent URIs are built from our knowledge of the domains we work in…….

BBC MMVIII

/Programmes Domain

BBC MMVIII

Domain modelling tips:

• You’re modelling things not web pages.• Trying to understand the relationships

between the things.• I As work with domain experts to model the

domain and understand it.• Its hard, it takes time to get something you

believe in, but it’s a really really important step in semantic web design.

BBC MMVIII

Consider the Episode…

Can have a parent brand or series……

Episode 4 of Top Gear series 10 for example……

What is the URI?

BBC MMVIII

URI of Top Gear Series 10 episode 4.

• Could be: bbc.co.uk/programmes/topgear/series10/episode4.shtml

• But we don’t put our technology (.shtml) into the URI.

BBC MMVIII

We don’t put technology into the URI

• If users want the document, they don’t mind what technology delivers it, they just want the information at that URI.

• They don’t want to have to go to different URIs on different devices. Why do they have to remember?

• Remember everyware? Multiplatform is going to become more of an issue.

• The technology used might change over the next few years.

BBC MMVIII

URI of Top Gear Series 10 episode 4 ctd.

Could be:

bbc.co.uk/topgear/series10/episode4

But from our discussions with the domain masters (editorial staff).– The same episode could be broadcast by

different brands (Top Gear, Best of Top Gear, Top Gear Take 2 etc).

– The episode could be reused across different series.

– The episode could change its title over its lifetime.

BBC MMVIII


Oh dear.

BBC MMVIII


In fact the URI of the document about this episode is:

http://www.bbc.co.uk/programmes/b0087f0s

BBC MMVIII

BBC MMVIII

URI design and change


We’ve taken out everything that can change. We’re using our own identifier (http://www.bbc.co.uk/programmes/b0087f0s

) to label it because there wasn’t a suitable webscale identifier already in semantic web for programmes, and because everything else (title, series, brand etc) could change.

BBC MMVIII

URI design and change continued.


If it can change (brand, title, series, episode number), we’ve taken it out of the URI.

BBC MMVIII

Categories, taxonomys on the page….not in the URI.

BBC MMVIII

Categories, taxonomys on the page….not in the URI.

Because categories and taxonomies will change during the lifetime of the product and your website……

BBC MMVIII

URI design to maximise persistance.

Know your domain. Work out your business rules.

Think about all the things that might change (including each time you redesign the website).

Take them out of the URI.

www.mycompany.com/products/:product

Might be a good place to start, where ‘:product’ is an identifier.

BBC MMVIII

Designing for multiplatform

So, how to deliver the document across multiple platforms, at the same URI?

BBC MMVIII

Content negotiation example:

BBC MMVIII

Multiplatform design at the same URI:

• Delivering the same information but in the most appropriate way for that device.

• So the RDF document is just the same as the HTML document, but expressed in a different language.

• The mobile document is the same as the HTML, but perhaps without images, or with the current information emphasised and other information linked to rather than put on the page.

BBC MMVIII

bbc.co.uk/programmes pages are available in:

• HTML, RSS, Atom, JSON, RDF, ‘Mobile phone’ representations.

• Some of these formats are still ‘Beta’ as its something we’re still working on.

• In fact we can’t do content negotiation on /programmes yet, but we’re planning on doing it, and building our URIs according to that plan.

BBC MMVIII

But lets not worry too much:

• It’s a software engineer thing. • So I hope its enough that we know about it.• Each representation can be designed to

make the best use of that device.

BBC MMVIII

Even if we’re not dynamically publishing.

• We can build persistent URIs, by taking out anything that might change.

• Then when the site is dynamically served, and can handle multiplatform, we can swap in the dynamically served site at the same URIs.

• Then we can think about serving different representations for different devices at the same URI.

BBC MMVIII

Some closing thoughts

Think multiplatform, not just web for the PC, and mobile phone but all sorts of ‘everyware’.

BBC MMVIII

Mobile traffic growing quickly

There has been a 15% growth in BBC Mobile since October - now have almost 4m UK users

[BBC online reaches 24 million users]

BBC MMVIII

Ubiquitous User Experience

The best user experiences are integrated at every touch point.

Big business opportunity?

BBC MMVIII

Open your data - stake your claim…..

Even if you decide not to build a UX others on the web may be able to do something with your data.

Semantic web technologies like RDF offer standard ways to publish data.

BBC MMVIII

Persistence is king……

Above all make your URIs persistent, for your company, for google, and for semantic web.

BBC MMVIII

Thanks!

Many thanks to all the /programmes and /music teams, and the IA team:

Particularly Michael Smethurst,Deanna Marbeck and Silver Oliver.

Many many others…….too many to mention.

BBC MMVIII

Contact:

[email protected]

[email protected]

Questions?