Digital Object Prototypes

1. 10th European Conference on Research and Advanced Technology for Digital Libraries Digital Object Prototypes An Effective Realization of Digital Object Types K. Saidis1, G. Pyrounakis2, M. Nikolaidou2, A. Delis1 1Department of Informatics And Telecommunications 2Libraries Computer Center University of Athens

2. Outline • The problem: • Develop an integrated web-based DL that manages diverse & heterogeneous digital material in a unified manner • Our proposed solution: • Digital Object Prototypes – an effective realization of Digital Object (DO) types • DOPs Overview • Discussion

3. Pergamos: Un. of Athens DL • Over 1 million objects originating from 8 disparate collections • Heterogeneous, digitised material requiring detailed cataloguing • Motivation • Simplify & speed up cataloguing • Decrease development time • Avoid custom coding for handling variations of digital material

4. Digital Objects • Human generated artifacts comprised of 4 components: • metadata, digital content, relationships & behaviors • Several ways to encode them: METS, FOXML, RDF, MPEG-21 • Diverse kinds of material are represented using variations of DO components

5. The problem • DOs (e.g. books, papers, photos etc) should be treated in a manner that reflects the nature of the underlying “real-world” objects • DL systems do not capture, express or use the nature of DOs effectively: DL users are forced to deal with each object’s idiosyncrasies manually

6. Digital Object Prototypes • Introduced in last year’s ECDL • A Digital Object Prototype (DOP) is a DO Type specification, providing a detailed definition of a DO’s constituent parts: metadata, files, relationships and behaviors • With the use of DOPs, high level services can handle DO component variations in a uniform manner

7. DOPs Contd. • DOPs provide the means to generate user-defined types of digital objects • DO type conformance: DOs automatically conform to their respective DOP • Inexpensive to add new types of digital material or modify existing ones (no custom coding required)

8. DOPs Contd. • DOPs draw on the notions of the OO model to implement DO types • All objects have a type • Objects honor encapsulation • Inheritance (under development) • DOs are conceived as instances of their respective prototype • Stored DOs are treated as serializations of DO instances

9. A DO Instance

10. Instantiation / Serialization

11. DOPs 3-tier DL Architecture Composition of DO behaviors DO Typing & Instantiation Storage

12. A DOP specifies • Metadata element sets • File types / conversions / batch ingestions • Permitted relationships • Private & Public behaviors

13. Behaviors • Encapsulation: Private & Public behaviors • Private behaviors are executed by the DO instance “behind the scenes” (the object itself triggers selected functionality) • Public behaviors are exposed to clients (clients trigger selected functionality)

14. Metadata Specification • DOPs specify: • the metadata element sets used by all objects of this type • the individual elements (fields) that constitute each set • possible mappings among elements of these sets

15. Metadata Specification II • Automatic loading / serialization of metadata sets • Behavioral characteristics of metadata elements: • isMandatory, isHidden, isRepeatable, defaultValue, validation • facilitate services to adjust to the object’s requirements • Mappings are automatically applied by the DO instance

16. Files Specification • DOPs specify: • The permitted files and their allowed formats • The information required to convert files from one format to another • Batch file ingestion / automatic DO creation process for each file

17. Example • Historical Archive’s Senate Session Proceedings > Folders > Sessions > Pages • What is a Page DO: • A digital image stored in various formats (e.g. high quality TIFF, web quality JPEG, thumbnail) • What is a Session DO: • A complex DO, consisting of Page DOs and session proceedings metadata

18. Example Contd. • Automatic file conversions • Page DOs automatically generate the two derivate formats from the high quality image • Batch Content Ingestion / DO Creation • Session DOs generate Page DOs from zip files containing high quality images

19. Session DOP

20. Page DOP

21. Relationships • DOPs specify: • The structural relationships the objects are allowed to participate in (e.g. child-of, or is-parent-of) • Whether the objects are allowed to reference other entities, either internal (residing in the same DL) or external ones

22. DOPs and Collections • DOPs “live” in the context of a collection (collection pertinent scope) • Collections are represented in terms of “special-purpose” DOs • Collections may contain other collections (hierarchical namespace) • The Digital Library is the collection of all collections

23. Pergamos Collections & DOPs

24. A hierarchy of DO instances

25. dl.histarch.* DOPs

26. dl.folklore.* DOPs

27. dl.papyri.* DOPs

28. dl.theatre.* DOPs

29. Summary • We argue that digital objects do have types (they fall into categories with uniform characteristics and behavior) • First realization of DO Type conformance: DOs honor the specifications residing in their DOP automatically • Effective management of diverse digital material: the DL system interprets DO types, not the user

30. Future Work • Implementation of DOPs inheritance (reuse: new types can be derived from existing ones) • Offline DOP validation process • Make DOPs operate independently of the underlying repository • Open source release of the DOPs framework

31. Thank you • Questions? • Email: saiko@di.uoa.gr • DOPs in action: Pergamos demo tomorrow (Poster #17)