Ontology

Wikipedia, Wiktionary, and sister projects do not have any explicit ontology—yet. They do not even have a defining vocabulary—yet. Ideally they would not reinvent the wheel but make use of a foundation ontology such as that created by the IEEE—extending that by some means approved by the board eventually.

It's a long way from a few tags to a proper tag schema to a database schema that's comprehensive, and ultimately a real ontology as such.

Ontological issues that plague Wikipedia include:

1. those already part of the Wikipedia syntax and DTD
2. those proposed for Wikitax, a future markup standard
3. those required to free Wikipedia3 from dependence on wikipedia.org and .com
4. those required for Wikipedia4 features - not defined yet, but likely to be well on the way by the time that any real ontology is being used.
5. (add more)

• Question: If Wikipedia proved so succesfull so far (without any explicit ontology and any complex explicit semantics behind it), what added value would this have? What is the trade-off between complexity and ease-of-use for a large number of editors, which may not have any technical background on this (but which are surely needed to continue to support the project)? Moreover, I personally find text-based search of content worked quite satisfactory so far. This is not a criticism of the idea, merely a point of discussion.

• One Answer: We can view the various foundation wikis (Wikipedia, Wiktionary, ...) the same way that the nine blind men viewed the elephant, each from a different point of view. What is needed is a way to make that point of view known, not a way to force the great mass of reality into line with a particular conceptual framework. Perhaps it is best that, where a sufficiently large community exists with a sufficiently high level of expertise, that community should create a conceptual framework and undertake to categorize or tag the entire content of the various wikis within that framework. This could exist in a separate realm from the materials being organized (ontologized?) just as our personal conceptual framework exists apart from the reality that it references. There would be strong constraints and disciplines on the things that could be entered into such a framework, since it would be encouraged to exist purely in reference to the wikis that it categorizes. It would be most interesting if each ontology effectively submerged the referenced sources within its own structure.

As an example there is the proposal to create a genealogy wiki (Wikifamily). One would have to grant that genealogy is one way to organize the human race. There could be great value in being able to collaborate in a public domain realm on this kind of research. Conversely, such a project has the potential to rapidly grow and get out of hand if allowed to carry extensive biographical data on every person who has ever (verifiably) lived and/or died. The solution is to restrict the structure to the minimum needed to reference and identify each person in the tree, focusing on links to other sources (including Wikipedia itself) to provide the great body of information. In this case, the community and the expertise already exist.

This approach also removes the requirement for increasing levels of complexity in the baseline Wikimedia software. There are probably countless conceptual/ontological frameworks that could add dimensions to the work of the Wikipedia Foundation. How all this relates to NPOV is debatable. Fortunately, NPOV does not mean "No Point of View".