Presentation on theme: "Content model workstream phase 1. Business transformation goals Efficient content updating and publishing New business opportunities enabled Efficient."— Presentation transcript:
content model workstream phase 1
Business transformation goals Efficient content updating and publishing New business opportunities enabled Efficient location of internal content Fine-grained business analytics
Phases of the content modelling workstream phase 1: creating the model phase 2: testing and refining the model phase 3: handing over the model to DK phase 4: migrating content to the new model
Phase 1 deliverables A model of how DK content can be structured as content objects Ontologies, which support content search and reuse A prototype to demonstrate content objects and ontologies
What are content objects? Reusable "chunks" of meaningful content, which can be combined to created products
DocBook Assembly DocBook’s method for topic-based authoring. Can be chunked at any level (whole chapter or a paragraph) Can be transformed from or to a traditional book structure
DocBook Toolkit Out-of-the-box production of PDF, HTML, EPUB, slides etc (good for proofing) Excellent tools support (Oxygen XML) Fully customizable - can layer own XSL on top Disassemble existing DK-Schema books using standard XSL
Introduction to linked data
Linked data is about facts Fact: The national gallery is an art gallery Fact: The national gallery is in Trafalgar Square Fact: Trafalgar Square's nearest tube is Charing Cross
Very simple facts Thing... has some property... value subject... predicate... object
Facts are stored in a data- base called a triplestore Ask questions like – what cultural buildings have a nearest tube station of Charing Cross?
Linked data uses the web nster Linked data stores are schema-less Got a new fact – just dump it in the store
What facts might be useful to DK?
Facts about real-world things For example, the National Gallery - type, nearest tube, lat/long, wheelchair - access, opening times, tours, events... Usually called "reference data" Store them once – accessible to all titles and products
Facts about content this content object is about the national gallery, was written by..., uses this image..., which was taken by..., which is rights cleared for the UK, is on page... of the 2012 edition of the RG to London
Facts need a vocabulary That's what the ontology is A vocabulary for writing down facts The ontology is divided into a number of separate modules
There are currently ontology modules assembly asset attraction book content location product transport travel-product travel web
There are extensible controlled vocabularies book-categories brands series travel-content travel-themes
Meeting the business transformation goals Efficient content updating and publishing - Moving towards a "create once, publish everywhere" strategy, by separating content, structure, meaning and presentation New business opportunities enabled - Leverage other sources of data - Increased brand awareness
Meeting the business transformation goals Efficient location of internal content - Use linked data to deeply query our content - Quickly respond to question like "how much content is there about hotels in Paris?" Fine-grained business analytics - Find out how content is performing at the object level - Could inform commissioning/ editorial decisions
Next steps Requirements from February 18th what does the content model need to do? Phase 2 throughout March Creating products from content objects Searching for content objects Modelling more domains Phase 3 beginning in April? Handover to DK Migration beginning in April?