Transcript

York Digital Library – Images Case Study

Julie Allinson, SUETr Multimedia Deposits Event30th March 2009

Overview

• A walk-through of what we did at York,

• bearing in mind– starting from scratch– we use ‘F’-for-flexible Fedora– and Mura front-end– short-term needs and long-term plans– not just an image library

Approach

• Image Content Model– a statement of requirements– a record of decisions– a design blueprint– [not yet] a machine-readable ‘content

model’– re-usable– possibly of wider use?

https://vle.york.ac.uk/bbcswebdav/xid-316149_3

Considerations

• Use case(s)• Image types• Uniqueness (preservation/access)• Metadata for intended and potential usage• File formats, sizes, expected delivery

mechanism• Access Restrictions• Deposit Workflow• Rights

Use Case(s)

• The University of York History of Art Department need digital images for teaching and research. They have no networked way of sharing image resources. They use a lot of scanned images under CLA, which need access control. They also sometimes take images and in the past got slides made of these.

But, other types of image too …

• Works of art• Architecture or architectural features• Archaeological sites, objects• Archival materials• Computer-generated images• Photographs taken for specific

purposes

Metadata

• Requirements• Model

– based on VRA – image & work distinction

• Profile– of VRA, minimal extensions

• Vocabularies• IDs and Labels• Mapping to DC

Datastreams/files

• Image formats– accept all– Preferred: open formats– Convert to: JPEG/TIFF for delivery

• Images ‘expressions’ stored– ARCHIVAL_IMAGE; FULL_IMAGE;

PREVIEW_IMAGE; THUMBNAIL

• File sizes and resolutions– maintain original size/resolution - guidelines

available [soon]

Relationships

• Important in Fedora – RELS-EXT• Currently stores collection

membership• Can be exploited further, e.g. to

store relationships between different objects, formal content model

• Exploring RELS-INT for internal relationships

Workflow

• Specification of the deposit process• Currently fairly simple• Uses

– mapped drives (local deposit)– muradora– xforms

• Has user-facing and technical aspect• Review and publish stage

Workflow – human

• Select object type• Select collection• Drop + select file• Describe• Save• Review – record + data dictionary• Publish

Access Control

• Requirements for access control• Two main current use cases

– Restricted CLA– Restricted public

• This needs– Integration with LDAP/Shibboleth– Access control filter to ‘talk’ to different systems– Roles and role mapping– RELS-INT datastream

Rights

• Copyright clearance info built into VRA• Standard rights statements – CLA &

Creative Commons built into VRA

• Image license for accepting collections

• Uses Creative Commons

It’s all a work in progress

• To-do– JHOVE for technical

metadata– Refining/extending

workflow– Access control– Different types of

images– etc. etc.

Usefulness of ‘Content Model’

• ‘checklist’• might be too internal to be of wider use• whether documented or not, each

aspect of the process has be thought through (with Fedora anyhow)

• a document helps demonstrate all the work involved

• needs regular review to keep up-to-date

Experience with images

• Good starting point– one dimensional– lots of work/advice– tools and code available– lots of content available– relatively small file sizes– easy to demonstrate

That’s it

contact ja546@york.ac.uk

top related