Use Cases and Requirements document resource page
This page is intended to support collaborative development of a Use Cases and Requirements Document for OWL 2
The goal of the requirements document is to capture the needs of user and implementer communities that motivated this revision to OWL.
Links suggested by Bijan:
- Feature categorization of motivations from OWLED2007 papers
- OWL1.1 Overview
- Michel's OWLED 2008 DC paper on modeling life science knowledge
- Next steps for OWL from OWLED2006
- Bijan's spreadsheet of OWLED2007 papers and features requested
- Weblog entry with quotes from OWLED2007 on features in OWL1.1
- short IBM report on a few uses of OWL
- OWLED2006 paper evaluating how OWL was used by modelers
In addition to the requirements motivating the language features, add material about motivation for changes in the abstract syntax and addition of the functional syntax.
- Problems with OWL Syntax
- DIG reference Middleware
- Igniting the OWL 1.1 Touch paper, the OWL API
- Supporting Early Adoption of OWL 1.1 with Protege-OWL and FaCT++
- The Protege-OWL Experience (some stuff on datatypes)
- Dan Connolly email on OWL next and RDF
- email on practices for defining properties and effects on cardinality constraints
- Holger's email querying for use cases for new meta properties and SelfRestriction
- AlanR question about issues with rdf mapping for annotations
- Dave Turner question on OWL2 support for assertions for anonymous individuals
Christine Golbreich's OWLED document on examples and use cases for OWL1.1 features (oct 2007 working draft):
and other links collected by users in LS and Industry (oct 2007)
- Requirements INSERM Hospital of Rennes
- Requirements Hospital Freiburg
- Requirements Agence de Biomedecine
- Requirements molecular knowledge base
- Requirements gSEM
- Semantic Web Rules for HCLS
- Labeling Brain Images
- Annotating brain MRI images
- Requirements Mitre
- Requirements Renault
Vipul Kashyap's proposal for Use Cases and Requirements Document Framework to motivate OWL 2.0 Design:
- Issue 124, discusses that ComplementOf applied to data ranges includes all datatypes not in the datarange, which is not very useful for modeling. Justification for having it in the language is based on Boris explanation that this form of ComplementOf comes up in common logical tranformations, and there is a desire to be able to communicate such transformed ontologies between reasoners.
- The addition of the reverse RDF mapping is motivated by the fact that this was simply missing in the previous specification and therefore difficult for developers to implement.
- There is general desire to keep the vocabulary small, which partially motivated the removal of object/datatype property punning.
- Regarding profiles, there were concerns that the practical difference between OWL Lite and OWL DL was small, motivating new fragments that had more clear benefits. On the other hand, it was thought that too many profiles in the language could overwhelm the developer community (there already being a concern by some that OWL is "too complicated"), motivating the choice of fewer profiles than initially proposed.
- Structural equivalence and structural consistency are motivated by implementor desire to have a clear "object oriented" model for what an ontology is.
- Other types of requirement categories:
- Semantic Web interoperability
Include a major section oriented to the viewpoints of specialized disciplines such as Life Science, Telecommunications, Manufacturing, and Earth and Space Sciences. Describe modeling problems/use cases from each of these domains and the language features they require. Map these into the appropriate place in another major section which more generally describes the language features and links these into more detailed descriptions in other OWL2 documentation.
New Features and Rationale (Working draft of the Requirements)