are machine-usable knowledge bases designed to sustainably change and evolve. They are ideally suited to a variety of tasks where it is impossible or undesirable to specify and enforce a rigid ontology, letting the ontology grow and evolve while while sustaining its basic functions. This includes managing human-generated metadata, fast prototyping of semantic applications, and coordination of distributed team activities. Organic ontologies typically consist of a rich diversity of disambiguated terms connected by a web of precise relationships and general associations.
Organic ontologies balance idiosyncracy (the multiplication of descriptive terms by individuals and communities) with inference (the automatic connection of distinct terms with one another). Properly balanced, inference reduces the chaos of idiosyncracy while idiosyncracy increases the reliability and precision of inferences.
The underlying mantra of organic ontologies is extend but connect, encouraging the creation of new terms while emphasizing or requiring their connection to existing terms. This strategy can be useful in both social contexts, where a diverse community is using a shared language, and in engineering contexts, where a small community is creating programs and processes which interact with the real world.
Communities of interest need languages for describing what they are talking about, inventing, or discovering. An organic ontology allows the community language to grow and evolve without degrading as the community grows or the sphere of discussion enlarges. Unlike folksonomies , which use social pressure to (imperfectly) force coherence and consistency, organic ontologies connect different terms and usages across individuals and communities. Organic ontologies require a little more work, but technology and the ontology itself can lessen the load. And the alternative, "ontological mob rule," may both trample important distinctions and scale into confusion.
Problem-oriented engineering can benefit from organic ontologies in the inevitable cases where a problem expands, changes, or becomes better understood. An organic ontology, by encouraging the specification of new terms to describe new concepts or distinctions, avoids the technological equivalent of putting new wine in old bottles, keeping both the old and the new alongside one another. Because organic ontologies are designed to change, the initial "ontology bottleneck" for project startup is reduced and projects can immediately start creating knowledge and building prototypes, expecting the knowledge to evolve as they understand or explore the problem domain further.
Technologically, organic ontologies require flexible platforms and work best with dynamic languages. In the language of conventional databases, organic ontologies involve onging schema changes (to handle idiosyncracy) and pointer intensive queries (to follow the connections that link terms to one another).