This document discusses research objects (ROs), which aggregate resources to bundle the contents of a research work. ROs allow for process preservation, reusability, repeatability, traceability, understandability, and curation of research. The RO model uses the Open Annotation vocabulary and can be extended for different domains. An example of a life sciences RO is provided. Creating ROs involves assigning permanent identifiers to resources, describing relationships between resources, and adding descriptions to a manifest file. Tools exist to help with RO creation and annotation, and ROs can be shared through repositories like RO Portal and ROHub.
Separation of Lanthanides/ Lanthanides and Actinides
Research Objects in Scientific Publications
1. Research Objects in Scientific
Publications
Daniel Garijo
Ontology Engineering Group
Universidad Politécnica de Madrid
Madrid, Spain
dgarijo@fi.upm.es
OEG, October 17th 2013
2. Outline
•
•
•
•
•
What is a Research Object?
Why Research Objects?
Research Object model at 5000 ft.
An example from the Life Sciences.
Research Object Overview.
• ORE
• Annotations
• Extensions
•
•
•
•
•
Eating our own dog food
6 step example
Cost vs benefit
RO Portal
RO hub
Research Objects in Scientific Publications
2
3. What is a Research Object (RO)?
Aggregation of resources that bundles together the contents of a research work
Research Objects in Scientific Publications
3
4. Why Research Objects?
•
Process preservation: URIs/DOIs for referencing resources.
•
Reusability of any part of the RO
•
Repeatability /Reproducibility: redeployment of the method
•
Traceability and error detection.
•
Attribution: able to cite data and publications of the RO
•
Understandability: Links between data, results and annotations.
•
Curation: by explicitly exposing the methods of the experiment.
Research Objects in Scientific Publications
4
5. RO Model at 5000 ft
+ Open Annotation
Research Objects in Scientific Publications
5
7. Research Objects: An Overview
•Vocabulary for describing Research Objects
•Generic
•Extensible to multiple domains
•Modular
RO Core
AO/OA
Research Objects in Scientific Publications
ORE
7
8. The Research Object Model: ORE
•ORE: Object Reuse and Exchange
•Resources can be further specialized according to the domain
Research Objects in Scientific Publications
8
9. The Research Object Model: Annotations
Research Objects in Scientific Publications
9
10. The RO Model
•Additional modules can be described on top of the Core
•Evolution
•Workflows
•Publications
•Etc.
Workflow
Evolution
Wfprov
Other
Wfdesc
RO Core
AO/OA
Research Objects in Scientific Publications
ORE
10
11. Eating our own dog food
•
Reproducibility is not only an issue
in the Life Sciences.
•
In computer science we:
•
Run experimental evaluations.
•
Reuse previous algorithms
and tools.
•
Build pipelines for processing
the inputs and obtain our
results.
http://petfoodadvisor.net/wp-content/uploads/2013/02/Cheap-Dog-Food.jpg
•WHY DON’T WE USE RESEARCH OBJECTS?
Research Objects in Scientific Publications
11
12. How? Minimal step by step example
• When writing a paper about a tool/algorithm/evaluation/survey
1. Create a permanent URL (purl, DOI, etc) for the Research Object
2. Distinguish between the resources that have been used from those that have
been generated.
•
A result was generated using/testing an algorithm/tool/framework.
•
An ontology was created from the CQs and ORSD of the domain.
•
A survey was performed by looking at these papers in the bibliography.
3. Describe the relationships among the resources and their dependencies.
4. Describe the resources in the RO (AO/OA, DC, schema, FOAF)
5. Add all the descriptions to a file (manifest)
6. State that the manifest is describing the purl of the RO.
7. You are done.
Research Objects in Scientific Publications
12
13. Annotating RDF/RDF-a can be exhausting
• Cost vs benefit.
•
It takes time, BUT
•
All your resources become easily
referenceable:
•
You provide context to your research
•
You ease the task of reusing your
software/papers
•
•
http://i3.kym-cdn.com/entries/icons/original/000/006/725/desk%20flip.jpg
There are tools to ease the task.
You don’t need to annotate everything!
•
The more you annotate, the better
your results will be understood.
Research Objects in Scientific Publications
13
15. rohub.linkeddata.es
•Many approaches exist for making your stuff persistent
•Figshare, Purl, Datacite, RO Portal, etc.
•Rohub.linkeddata.es is a page (under development) for
pointing at the resources created in ROs of the OEG.
•3 ROs at the moment.
•Content negotiation (TTL and more)
•Annotated in RDF-a
•ROs as web pages with purls!
•http://rohub.linkeddata.es/ro-svmworkflow/
(http://purl.org/net/svm-opt-research-object)
http://esp.habitants.org/campana_cero_desalojos/jornadas_mu
ndiales_cero_desalojos__por_el_derecho_al_habitat_2013/work_in_progress
Research Objects in Scientific Publications
15
16. RO-ify your work!
http://researchobject.org/
Join the discussion…
W3C Research Object for Scholarly Communication
Community Group
http://www.w3.org/community/rosc/
Acknowledgements: Wf4Ever project
17. Research Objects in Scientific
Publications
Daniel Garijo
Ontology Engineering Group
Universidad Politécnica de Madrid
Madrid, Spain
dgarijo@fi.upm.es
OEG, October 17th 2013