PerfectO: An Online Toolkit for Improving Quality …
181
5.2 Semantic Web Community Evaluation Criteria
To leverage and identify value for our PerfectO methodology, we discuss the evaluation criteria as outlined by the International Semantic Web Conference (ISWC)
conference and their call for resources (e.g., ontologies) track papers [54].
Impact Criteria. PerfectO has a significant impact outside of the semantic web
community. It has been mainly applied to the IoT community but is generic enough
to be applied in other domains. As demonstrated in Sect. 4.2, PerfectO has been
used with the LOV4IoT ontology catalog which covers more than 20 applicative
domains (introduced in Sect. 2.1.1). It is the first innovative platform for semantic
interoperability and disseminates best practices and existing tools as a set of catalogs
following “Knowledge Directory Services” approaches. Our first step is to focus
on ontology interoperability and improvement. Instead of saying that our tool is
better than existing ones, we exploit their potential when they already provide web
services. Our literature review shared in a structured way on the web through FAQs
and mindmaps explained in Sect. 6 is also useful for teaching, tutorials, etc. We
have in mind to have a similar approach for datasets, reasoning, and querying, as
explained in Sect. 4.1. We are aware that researchers can consider our methodology
as not innovative since we reuse and integrate tools, but we are following software
engineering research methodologies [14].
Availability Criteria. The PerfectO software is under the GNU GPLv3 license.
A GitHub repository has been initiated https://github.com/perfectkb/perfecto which
stands for perfect knowledge bases. The LOV4IoT catalog has been designed to
extend the LOV ontology catalog since LOV inserts only new ontologies when best
practices are followed. Due to this experience, we realized that the learning curve to
improve ontologies and use the tools needs to be reduced and disseminated within
other communities (e.g., IoT community). This work is aligned with our current
activities at improving ontology quality in any projects designed at the Knoesis
research center until 2020. PerfectO is hosted on Google Application Engine to
avoid any server and DNS maintenance related issues.
Reusability Criteria. We provide a “Semantic Web Best Practices for Dummies”
documentation [88] to improve ontologies. More than 200 ontology URLs have
been referenced with the LOV4IoT dataset which has been integrated with PerfectO.
PerfectO can be applied to any ontology URL. It provides online tools; no set up is
required which reduces the development time. It can be extended with more tools to
improve ontologies. For instance, we can integrate ontology matching tools such as
LogMap, and ontology catalogs such as BioPortal and LOV to automatically suggest
a new ontology. We are aware of the current limitations of PerfectO: some tools
cannot load some ontologies. We try to understand why to help beginners solve the
issues and report the solutions within the documentation.
Design and Technical Quality Criteria. We classify, reference, synthesize, vulgarize and disseminate best practices learned from the semantic web community to
other communities (e.g., IoT). We reuse existing tools for automatic visualization,
documentation and ontology quality as explained in Sect. 7.3.
181
5.2 Semantic Web Community Evaluation Criteria
To leverage and identify value for our PerfectO methodology, we discuss the evaluation criteria as outlined by the International Semantic Web Conference (ISWC)
conference and their call for resources (e.g., ontologies) track papers [54].
Impact Criteria. PerfectO has a significant impact outside of the semantic web
community. It has been mainly applied to the IoT community but is generic enough
to be applied in other domains. As demonstrated in Sect. 4.2, PerfectO has been
used with the LOV4IoT ontology catalog which covers more than 20 applicative
domains (introduced in Sect. 2.1.1). It is the first innovative platform for semantic
interoperability and disseminates best practices and existing tools as a set of catalogs
following “Knowledge Directory Services” approaches. Our first step is to focus
on ontology interoperability and improvement. Instead of saying that our tool is
better than existing ones, we exploit their potential when they already provide web
services. Our literature review shared in a structured way on the web through FAQs
and mindmaps explained in Sect. 6 is also useful for teaching, tutorials, etc. We
have in mind to have a similar approach for datasets, reasoning, and querying, as
explained in Sect. 4.1. We are aware that researchers can consider our methodology
as not innovative since we reuse and integrate tools, but we are following software
engineering research methodologies [14].
Availability Criteria. The PerfectO software is under the GNU GPLv3 license.
A GitHub repository has been initiated https://github.com/perfectkb/perfecto which
stands for perfect knowledge bases. The LOV4IoT catalog has been designed to
extend the LOV ontology catalog since LOV inserts only new ontologies when best
practices are followed. Due to this experience, we realized that the learning curve to
improve ontologies and use the tools needs to be reduced and disseminated within
other communities (e.g., IoT community). This work is aligned with our current
activities at improving ontology quality in any projects designed at the Knoesis
research center until 2020. PerfectO is hosted on Google Application Engine to
avoid any server and DNS maintenance related issues.
Reusability Criteria. We provide a “Semantic Web Best Practices for Dummies”
documentation [88] to improve ontologies. More than 200 ontology URLs have
been referenced with the LOV4IoT dataset which has been integrated with PerfectO.
PerfectO can be applied to any ontology URL. It provides online tools; no set up is
required which reduces the development time. It can be extended with more tools to
improve ontologies. For instance, we can integrate ontology matching tools such as
LogMap, and ontology catalogs such as BioPortal and LOV to automatically suggest
a new ontology. We are aware of the current limitations of PerfectO: some tools
cannot load some ontologies. We try to understand why to help beginners solve the
issues and report the solutions within the documentation.
Design and Technical Quality Criteria. We classify, reference, synthesize, vulgarize and disseminate best practices learned from the semantic web community to
other communities (e.g., IoT). We reuse existing tools for automatic visualization,
documentation and ontology quality as explained in Sect. 7.3.