OntologyDesignPatterns

Ontology design patterns

This page links to subsidiary pages that are used to gather use-cases and outlines for ontology design patterns. In due course, we expect this information to re-appear in a more formally published form, such as Mikel Egaña Aranguren's work at Manchester University on a design pattern catalogue [1].

Contents:

1. Adding a new pattern
2. Patterns and problem statements
3. List of pattern pages
4. References

1. Adding a new pattern

To add a new pattern or problem statement:

1. Edit this page to create a link to a new page describing the pattern. I suggest using page names of the form "Pattern/NameForPattern", so that a list of pattern pages can be automatically retrieved (see #List below).

2. On the updated version of this page, click the link (or nearby "?") just created, and you will be asked to create a new wiki page. A list of page templates are displayed: choose "PatternTemplate".

3. Fill in the pattern details in the new page thus created.

2. Patterns and problem statements

Note: (FwL) refers to patterns that were suggested by Freshwater Life at the 3rd Ontogenesis Network meeting.

• Pattern/ReifiedProperties - Reified property instances (FwL).

• Pattern/ReifiedEntities - Reified entities (a possible alternative to reified properties).

• Pattern/UserConceptLabelling - concept labelling to support reasoning and better user interfaces (FwL).

• Pattern/ConflictingTheories - Representing information in a single ontology store to allow for conflicting theories (FwL).

• Pattern/MereologicalRelationshipsRCC8 - RCC8 is a logical framework devised by researchers at Ordnance Survey for describing and reasoning over topological relationships between geographic features.

• Pattern/ObservationsAndConclusions - Inspired by work on an animal behaviour ontology at Oxford, describing behavioural characteristics can raise issues of separating observations from interpretations (e.g. mastication vs feeding).

• Pattern/InstantsAndDuration - Temporal modelling commonly raises a need to distinguish between instants or periods of time and durations. Also, for some kinds of scheduling, time relative to a background cycle is important (time of day, or season of year).

  • Alan Rector mentioned Allen calculus: 13 relations between intervals. See reference [3].

• Pattern/PositionAndExtent - Spatial modelling commonly raises a need to distinguish between specific positions and measures such as length, area, etc.

• Pattern/PurportedTime - Some data sets need to distinguish between actual time when an event occurred, and simulated or depicted time (e.g. scientific experiments involving time-shifting

• Pattern/RepresentationAndReality - an issue mentioned by Ordnance Survey researchers is that they are dealing with two related modelling problems: describing geographic features in the real world, and describing large datasets that represent observations of those geographical features. These might be characterized as territories and maps.

• Pattern/ArbitraryBoundaries - there are many cases where one is describing a process or structure in which identifiable phases merge into each other (e.g. stages of frog development, river flows into lake), and given any particular observed instant it is not generally possible to determine its phase, or is possible only through consideration of the whole of which the phase is a part.

• Pattern/PhasesOfIndividual - how should developmental or other phases of an individual (e.g. stages of frog development) be modelled with respect to the corresponding individual?

• Pattern/PossiblyAndSometimes - some features are defined by transient conditions (e.g. a desert river, which may be dry for much of the year). How are such conditions logically described.

• Pattern/DefaultsAndExceptions - many logical frameworks depend on logical monotonicity (truths don't become falsehoods through the addition of new information), which is in conflict with default reasoning (a bird can fly ... unless it's a penguin).

• Pattern/NonOntologyDescriptions - ontologies are widely seen as the primary means of capturing knowledge in machine-processable form, but there are other techniques. What are these and when should then be used.

• Pattern/KnowledgeInDataSchema - vast amounts of knowledge are stored in existing databases, implicit in the database schema or other data structure used: how can we unlock this knowledge for more general use?

• Pattern/EvolvableOntologies - an ontology represents aspects of a state of knowledge about some domain; how can it evolve to take account of evolving knowledge, while retaining access to facts stated with respect to earlier art.

• Pattern/RequirementCreep - a reported problem when designing ontologies is that the original requirements are rapidly extended when an ontology comes into use. How can such requirements creep be managed? Can techniques akin to agile software development be applied?

• Pattern/OntologyModularization - how can ontologies be designed so that the different elements of domain knowledge can be selected and re-used to serve different target applications.

• Pattern/TerminologyUsage - Practice vs technical usage (e.g. anatomical part-whole as spatial or functional).

• Pattern/GlobalLocal- a socio-technical pattern addressing the tension between (i)the need for technical inter-operability(one size fits all), and (ii) the need for local or temporal felxibility (heterogeneous or evolving requirements, volatile environments)

3. List of pattern pages

4. References

1. http://www.gong.manchester.ac.uk/ontologydesignpatterns/ - Ontology Design Patterns (ODPs) Collection

2. http://www.scs.leeds.ac.uk/spacenet/publications.html - publications page of The Qualitative Spatial Reasoning Group at Leeds University

   • includes a 2001 survey article Qualitative Spatial Representation and Reasoning: An Overview which itself contains 180 or so references.

3. http://portal.acm.org/citation.cfm?id=358434, http://www.cs.iastate.edu/~honavar/allen-temporal.pdf - Maintaining knowledge about temporal intervals, James F. Allen, Communications of the ACM, Volume 26, Issue 11 (November 1983), pp 832-843

4. http://www.ihmc.us/users/phayes/Pub/TimeCatalog.pdf Pat Hayes, 1995, A Catalog of Temporal Theories, Tech report UIUC-BI-AI-96-01, University of Illinois

---

-- GrahamKlyne 2007-07-06 10:23:25