13
GRANT AGREEMENT: 601138 | SCHEME FP7 ICT 2011.4.3 Promoting and Enhancing Reuse of Information throughout the Content Lifecycle taking account of Evolving Semantics [Digital Preservation] “This project has received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no601138”. Encapsulation and digital ecosystem Anna-Grit Eggers (University of Goettingen) Efstratios Kontopoulos (CERTH)

PERICLES Encapsulation and Digital Ecosystem

Embed Size (px)

Citation preview

Page 1: PERICLES Encapsulation and Digital Ecosystem

GRANT AGREEMENT: 601138 | SCHEME FP7 ICT 2011.4.3 Promoting and Enhancing Reuse of Information throughout the Content Lifecycle taking account of Evolving Semantics [Digital Preservation]

“This project has received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no601138”.

Encapsulation and digital ecosystemAnna-Grit Eggers (University of Goettingen)Efstratios Kontopoulos (CERTH)

Page 2: PERICLES Encapsulation and Digital Ecosystem

Encapsulation Context

Introduction

Page 3: PERICLES Encapsulation and Digital Ecosystem

Objective1. Explore techniques to encapsulate digital ecosystem

information, in particular Significant Environment Information (SEI) and policy information.

2. Attach this information to the digital objects to which it refers to or for which it is relevant.

3. Describe considerations for creating components to support this objective.

4. Address a sheer curation scenario where information that is extracted from the ecosystem is encapsulated with the digital object to improve its reusability.

Page 4: PERICLES Encapsulation and Digital Ecosystem

Example: the PERICLES sheer curation scenario. The PERICLES Extraction Tool (PET) monitors the environment and reacts to changes in this environment. It extracts this information, and temporarily stores it locally. The extracted information is then appraised by the person creating and altering digital objects. The appraisal allows for intervention and filtering of problematic information, such as private or confidential information, not to be encapsulated.

Page 5: PERICLES Encapsulation and Digital Ecosystem

1. Extraction of significant environment information (SEI) and other relevant information (e.g. policy-related information)

2. Appraisal of the extracted information (what shall be encapsulated, what deleted?)

3. Aggregating digital objects and the appraised information.

Workflow

Benefit:• The packet of digital object and SEI can now leave the

environment at any time, e.g. by email or through a submission to a repository.

• No risk of loss of information necessary to improve the object’s reusability as it will also be available at the receiving new environment.

Page 6: PERICLES Encapsulation and Digital Ecosystem

● Payload = the actual ecosystem information● Which ecosystem information belongs to which digital

object?● Direct relation to digital object

◦ Metadata for example technical information about a picture, or descriptive metadata about a specific digital object

◦ File-specific SEI extracted by the PET tool or a policy that applies to a specific file or set of files

Embedding the payload

● Complex / ambiguous situation:◦ The payload referring to a collection of digital objects (e.g. a policy may

apply to a collection of digital objects)◦ Information relating to processing environment of a file (file-independent

SEI). ◦ Information that is not specific to a file or collection of files, e.g. parts or

the total of the ecosystem model. This situation could apply the structure and contents of the ecosystem model are important for the object interpretation and use, so that it is bound to the object.

Page 7: PERICLES Encapsulation and Digital Ecosystem

1. Embed related object or file metadata directly into the digital object

Embedding the payload (cont)

2. To avoid redundancy when embedding collection-related information:

◦ Embed the payload information into a subset of the digital objects

◦ Compute identifiers for each of the digital objects in question

◦ Embed the identifiers of the respective digital objects in the remaining digital objects of the collection.

◦ The embedded information can be recovered by accessing the identifiers of any of the elements of the collection, and from that accessing any of the digital objects.

◦ The use of the identifiers could be practical to refer to common (shared) information.

AVOID REDUDANCY

Page 8: PERICLES Encapsulation and Digital Ecosystem

3. Embedding payload information is not related to any specific file or collection, will require some technique to be able to find and make sense of such information in a collection of digital objects embedding it.

Embedding the payload (cont)

Example: information describing the ecosystem◦ For example, additional metadata files could be stored in the

collection, indication the nature of the embedded information and its location.

◦ It would be possible to store the ecosystem model as snippets by dividing the ecosystem entities, and to embed each entity into a different digital object for a collection.

◦ Given this information, the whole ecosystem could be reconstructed by decoding and merging the snippets stored in the collection of DOs.

Page 9: PERICLES Encapsulation and Digital Ecosystem

● An ecosystem model is expressed with a upper ontology (e.g. the LRM used in the PERICLES approach) and a domain ontology.

● Split the ecosystem model into smaller parts, e.g. by following method:◦ embed a representation of the digital object and their dependencies to

all other resources based on the upper ontology (e.g. LRM) in the form of metadata snippets accompanying the digital objects.

◦ these fragments are expressed by using either core constructs from the upper ontology (LRM) or the domain-specific ontology.

● This method of embedding fragmented information across multiple digital objects supports the reduction of redundancy of embedded information.

Encapsulating Ecosystem information

Page 10: PERICLES Encapsulation and Digital Ecosystem

Example: PERICLES scenario for embedding indirectly related complex information An LRM snippet containing a detailed DO description along with the representation of all associated dependencies can be moved between environments - i.e. in the figure, the DO initially lies in Environment A and is moved to Environment Z.

Page 11: PERICLES Encapsulation and Digital Ecosystem

● Ontology import is a key concept in ontology reuse, i.e. reusing constructs/axioms from an existing ontology when developing a new ontology.

Ontology import

● In OWL (the W3C Web Ontology Language) the main primitive for ontology import is owl:imports.

● Although ontology import displays certain non-trivial drawbacks (e.g. nesting of imports is typically not apparent, import is syntactic and not semantic, scalability issues etc.), it currently serves effectively the vision of a Semantic Web based on distributed ontologies.

Page 12: PERICLES Encapsulation and Digital Ecosystem

◦ Collect snippet from Environment A [encapsulate] → ship it to Environment Z [decapsulate].

◦ Collect snippet from Environment B [encapsulate] → ship it also to Environment Z [decapsulate].

◦ At Environment Z: Insert all snippets into an overall 'merged' ontology, to get the bigger view of the whole Digital Ecosystem (including all environments), and a view of the whole dependency graph.

Ontology snippet export and mergeSince a Digital Ecosystem may include manifold environments (i.e. machines), our proposal can also effectively tackle the following scenario:

Page 13: PERICLES Encapsulation and Digital Ecosystem

• Objective: encapsulate the SEI extracted from the PET tool into digital

objects• Solution:

a component called PET2LRM translating PET information into LRM

•performs a direct mapping of PET information into simple LRM constructs

•converts PET extracted environment information to LRM

PERICLES scenario

• Then embed the information using techniques proposed by PeriCAT.

• For not directly related information (events, Environment information from PET) can be accomplished using the techniques of spreading the information across the collection.