The EcoBuilder - A Tool to Integrate the Digital Ecosystem Model into the PERICLES picture
We are delighted to announce the release of the Digital Ecosystem Model (DEM) and the EcoBuilder, a Java tool to create DEM instances. The DEM is an ontology for modelling the interwoven relations, dependencies, and interactions of entities in digital ecosystems. Digital ecosystems encompass digital objects and their environments including their technical infrastructure, community relations, policies, and processes. The DEM provides also modelling concepts for the analysis of planned and unplanned changes of the underlying digital ecosystems. This blog post focusses on the EcoBuilder tool, while the the DEM will be discussed in a second publication. The DEM’s specification can also be found at the related deliverable D.5.3, which will be available later this year.
The EcoBuilder is basically a wrapper around the DEM that abstracts and facilitates its creation and integration with existing workflows. Each resource and relation of the DEM is wrapped through a Java class, which allows a less error-prone modelling on a more abstract level. The following small scale example shows the use of these template classes – it expresses that an important file lies on a server and can be accessed via a web interface:
The tool uses the Java Jena library to generate the ontology source code from the wrapper classes. Three different types of models can be generated with the tool:
1. The abstract DEM with all abstract entity templates and relations
2. Scenario models created by the user through the GUI, which encompass scenario specific instances of the abstract templates and custom templates
3. Scenario models created via the Java API, thus through the direct use of the Java template classes.
The EcoBuilder in a PERICLES workflow
A further purpose of the tool is the integration of the DEM into the PERICLES context, as well as into external workflows. It connects the DEM to other PERICLES ontologies and components. A wrapper of the Linked Resource Model is used by the EcoBuilder to create the DEM entities, as the DEM is built on top of the LRM. A modular design allows the integration of other ontologies into the EcoBuilder, which is demonstrated on the already integrated PERICLES Digital Video Artwork domain ontology. Modelling of preservation policies is supported by the EcoBuilder through the integration of SCAPE’s guidance policy elements.
The PERICLES project is reaching its conclusion and the connection of the various components, tools and concepts created by the project, viewed as a whole, deliver the PERICLES approach. The EcoBuilder integrates into PERICLES’s Model-driven preservation approach of ontology-aided preservation actions that are integrated into the environments in which digital objects are created and altered. It can be connected to the PERICLES Extraction Tool via mediator scripts to update the models with life information from the underlying ecosystem. Created DEMs can be sent to the PERICLES Entity Registry and Model Repository (ERMR) from where they can be used for further processing, e.g. to investigate ontology evolution, or the impact of model changes with the PERICLES MICE tool.
EcoBuilder at a glance:
- License: Apache v.2 Open Source license
- Systems: Linux, Windows
- Requirements: Java 8
- Interfaces: GUI and Java API
- Output format: OWL/XML and Turtle
- Download: https://github.com/pericles-project/EcoBuilder
Templates, Entities and Relations
We distinguish between abstract templates, as a “Digital Object” template, and concrete existing entities instantiated from these templates, as a “monthly report” that is an entity of the “Digital Object” template. Basic digital ecosystem templates are provided by the EcoBuilder, but it is also possible to create domain specific custom templates which inherit the principles of existing templates. Inheriting templates can use all relations which work for their base template. A custom “Project Report” template inherits from the “Digital Object” template, and can then be used as template for all future report entities. It can then automatically use the relation “hasAuthor” which is defined for the “Digital Object” template. Also domain specific relations can be created with the EcoBuilder.
Involvement of scenario experts
Ontologies are storehouses of knowledge that describe the world through the application of “Subject – Predicate – Object” triple sentences. Their flexibility allows the capture of complex coherences that would not fit into fixed data structures. Scenario experts are the ones with the knowledge about the digital ecosystems being important for preservation purposes and maintenance. They are not necessarily familiar with the RDF ontology language or Java. Therefore the EcoBuilder provides an intuitive graphical user interface that can be used to create custom templates, scenario entities, relations, and their descriptions. The ontology models are generated from this provided information.
This graphical interface reduces the complexity of the underlying ontology to a simpler level. The reduction has the great advantage that the modelling is practicable and that the results are reliable, but it also has the disadvantage that not all possibilities of ontologies can be exploited. An approach for applications that require the use of further ontology features is to introduce the missing principles into the generated models after the scenario experts created them. Another approach is the use of the EcoBuilder’s Java API in combination with Java Jena where all necessary features are available.
Different modelling strategies, that describe the level of detail and order of modelling, are discussed in the related deliverable.
Preservation and Analysis of Scenario Models
In most cases the created DEMs depicts only specific aspects of the underlying digital ecosystem, which we named the scenario. The scenario is modelled with a designated purpose in mind, but it can also be merged with further scenarios depicting other aspects of the system. Such an overall model of a digital ecosystem can later be reused for future scenarios.
A static scenario model can serve to preserve and explain specific knowledge of the digital ecosystem. It can also be analysed to identify dependencies, single point of failures, or the compliances of policies.
Scripts that use the EcoBuilder’s Java API can keep a model up to date. A script could for example be connected to a database to automatically create a digital object entity every time a new entry is submitted. Such an up-to-date model can be analysed to identify the impact of changes of the underlying digital ecosystem. Furthermore it can be used for change planning through the simulation of the planned changes. An analyse of the simulation can show the impact of the change before it is applied, and therewith help to prevent mistakes and to ensure the quality of the digital ecosystem.
A paper about the Digital Ecosystem Model and the EcoBuilder will be presented at the MEDES conference Nov. 2016 in Hendaye, France. The ontology and the tool will also be a topic at the PERICLES final event on 30 November - 1 December 2016.
It is planned to release two updates of the EcoBuilder before these conferences. The releases will include the ERMR connection and an update of the LRM wrapper to the newest LRM version. They will also address issues and feedback reported on GitHub.
Further material about the EcoBuilder, as a quick start and installation guide, will be published on GitHub and the theoretical background can be explored in the corresponding deliverables.
More information on the DEM will appear soon in a separate blog post. Watch this space!