Use Case

Bank intranet

• poorly organised content
• various levels of document quality -- which is the definitive?
• some obsolete documents
• loose vocab. for indexing
• mainly internal content (prod. docs; bus. proc. descriptions; client data; simulators)
• some docs from professional editors (legal/tax)

What are user needs?

• efficiently answering client ques.
• find docs. relevant to client needs
• document retrieval; current definitive copy
• personalisation of access depending on expertise
• capacity for users to build custom doc. organization

Organize Content

• Thesaurus

- bank terminology; glossary
- metadata repository
- linking related terms

Make Search Easier

- synonyms/acronyms
- different names for same operations in different depts.

• Knowledge Base

Next step is to create the knowledge base.

- Products: decomposition; relationships (semantic links); life cycle
- Bus. Processes: description of tasks by products and stage
- Bank organisation: departments, services

• Classifications, Taxonomies

Build classifications for:

- Product families
- Product functions
- Document types
- Client typologies
- Financial instrument

Biggest difference between classification and taxonomies is what you can charge

• Define scopes for personalisation

Expertise levels, etc.

Difficult to do, but should be simple. Not restriction of content, just providing simpler views in this use case.

What is Content Structure?

• Classifications
• Thesauri
• Topics and knowledge databases
• Bus. processes
• Doc. organizations

• Index "Reusable Content"

- at what level will it be accessed?
- define level of granularity of content to index (e.g. a task description, chapter)
- keep original document navigation (chapter nav.; index; summary)

Users were used to particular navigation elements.

• Create content structure repository

Classifications should be stable, otherwise users get lost/confused. Most changeable aspect is names (thesaurus).

Some content doesn't need 'managing' -- it's stable, and just needs storing (e.g.professionally produced content)

Need tools to work with the content structure repository: views; editing tools; import/export; APIs, etc

Which standard?

Topic Maps:

• powerful enough to describe any organization model (taxonomy, thesaurus, lexicon, knowledge base, document navigation)
• easy to manipulate and read
• scope mechanism for personalised publishing
• independent from content
• added feature: temporal validity of each object of the content structure

Which content structure model?

• build using existing org. model (reuse existing 'pre-XML' work) -- e.g. bus. process, doc. navigation; use standards
• build custom structure model for the knowledge semantic links (for particular client) -- products, company, clients, relationships

• Integration

-- Authoring Tools -- interact with content structure management system
-- Content Management System
-- Content Structure Mgt. System
-- Publishing

Results

Accurate and Efficient Access to Docs

• Filtering (using taxonomies)
• Search (topics, synonyms, product codes)
• Navigation (taxonomies, etc)

Dynamic Publication of Bus. Docs

• Personalized publishing
• PDF publication for clients
• Users able to build own doc. org to repetitive client needs
• Customized XML publication for the bank distribution channels

Conclusions

• content structure important
• keep your 'documentalists'
• talk to professional editors
• use topic maps!

Is a content structure repository a backbone for content intelligence tools? (open question)

• nat. lang. query
• meta tagging
• text mining
• auto. categ.
• auto. translation
• inference process
• inference engine

Impressive results from collating and publishing structure.

Questions: PeterP

• exposing content structure data as a resource?

M. B.

• not enough to buy CMS, or system: need human input to create repositories