Developing a digital repository infrastructure for Kings College London RSP Training Day, 22nd Janua

1
Developing a digital repository infrastructure
for Kings College LondonRSP Training Day, 22nd
January 2009

• Gareth Knight
• Centre for e-Research

2
Approach

• Analyse existing practices limitations of
  current system
• Establish requirements for Information management
  access
• Investigate alternative approaches (software
  choice, extensibility, applicability to your
  data, use by others)
• Prototype smaller projects and experiments

3
Centre for e-Research

• CeRch (http//www.kcl.ac.uk/iss/cerch) is
• A RD department in Information Services and
  Systems (ISS) that performs
• Management and preservation of research outputs
  from KCL researchers in all disciplines
• Research, teaching and consultancy on
  e-infrastructure, data curation and preservation
  and others.
• Formerly Arts Humanities Data Service
  Executive
• Management and preservation of research outputs
  from UK researchers in arts and humanities

4
Context Existing approach

• Formal, but manual ingest procedures
• Bespoke repository for data management
• Not scaleable code could not easily be
  reapplied to other projects.
• Functional limitations
• Preservation, provenance metadata
• Limited delivery systems
• Collection-level identifiers (mostly)
• Diverse, semi-structured data

5
Requirements

• Persistent identifiers down to the level of
  individual datastreams, accommodating compound
  content models
• Versioning of content and metadata
• Automated processing and user input
• Able to integrate specialised third-party tools
  (e.g. format conversion)
• Preservation metadata management
• Audit trail/provenance metadata
• Standard distribution methods for specific
  content types (Disseminators)

6
What do we use Fedora for?

• Digital repository
• Kings Research Archive An institutional
  repository for open access research papers
  written by Kings College London staff
• Virtual Research Environment (VRE) supporting
  research management
• EIDER Project Demonstrator for enhanced deposit
  and ingest
• Preservation Services
• SOAPI an architecture for (partially)
  automating preservation and ingest workflows in
  digital repositories
• SHERPA DP2 developing preservation services for
  content located in disparate locations.
• Digitisation projects
• Historical Hansard - Digitisation project
  scanning and markup of 50 years of debates from
  the Upper Chamber of the Northern Ireland
  Parliament from 1921 to 1972
• East London Theatre Archive - Digitisation of
  15,000 performing arts resources, from playbills
  and programmes to press cuttings and photographs
  from East London theatres

7
Capture Ingest workflow
Activities performed during Ingest
8
Metadata (1) Descriptive

• Each project has specific descriptive MD
  requirements
• Scholarly Works Application Profile (SWAP)
  created schema for IR
• Metadata Object Description Schema (MODS) ELTA
  and SHERPA DP2
• MarcXML SHERPA DP2
• Simple DC (various)

9
Metadata (2) SWAP
10
Metadata (3) Preservation

• Preservation
• PREMIS Object
• PREMIS Event (forthcoming)
• Generated by DROID, JHOVE others
• Rights
• Rights MD
• Provided by Sherpa-Romeo

11
Metadata (4) Preservation
Rights metadata provided by Sherpa Romeo
Technical metadata provided by JHOVE
12
Data Capture (1)Kings research data

• Collection of Kings research data
• Web interface for deposit
• Deposit via SWORD from desktop/web client
• Capture of metadata from Research Gateway, Web of
  Science and other sources.

13
Data Capture (2)Archiving services

• SHERPA DP2 provides archiving and preservation
  services for varied software repositories and web
  resources
• Content providers supported
• Repositories Fedora, CDS Invenio, DSpace,
  EPrints, DigiTool
• Website Large dynamic sites (through
  Subversion), static sites.
• Capture methods
• OAI-PMH for metadata capture
• Data capture over HTTP/FTP and VPN.

14
Digitisation (1)East London Theatre Archive

• 15,000 digital objects playbills, programmes,
  press cuttings and photographs.
• Object model representing 2 layers
• Performance venue
• Item (3 manifestations of each image
  (high-quality, distribution, thumbnail)
• Each will contain MODS metadata
• Accessible through browse, search Google
  maps-style UI

15
Digitisation (2)Historical Hansard

• 50 years of debates from the Upper Chamber of the
  Northern Ireland Parliament from 1921 to 1972.
• Separated into collection and volume.
• 45,100 items containing
• Page images (3 manifestations of each image
  (high-quality, distribution, thumbnail)
• OCRd text stored as XML
• Relationship MD
• UI Experiment with Fez, Muradora, Vital,
  Existing Stormont

16
Lessons we have learnt

• Understand your needs
• No one-size-fits-all approach
• Match requirements to functionality, not visa
  versa
• Implementation of a Fedora repository requires
  time
• No out-of-box solution, though likely to change
  in the near future
• Consider a long-term development plan. Some
  customisation may be required
• Consider future expansion plans
• Where do you want to be tomorrow?
• Dont be intimidated
• Lots of features, but dont need to use them all
• Possible to break implementation into
  well-defined stages
• Avoid reinventing the wheel
• Examine existing Fedora projects that may save
  development time.
• Develop code that can be repurposed to other
  project

17
Thank you!

• Gareth Knight
• Centre for e-Research
• gareth.knight_at_kcl.ac.uk