| ID |
Question |
Name |
| Shangguan Managing Ontologies GTW1 |
Does the OI model architecture not support functional and inverrse functional properties? These (IFPs in particular) are crucial in the use of distributed ontologies and data, but seem to be missing from this model. |
Gregory Todd Williams |
| Shangguan Managing Ontologies Jesse Weaver |
In the introduction, the authors give the example of a service provider B who wants to create a bicycle ontology (BO) reusing as many definitions as possible from an existing sports ontology (SO). The authors state that changes to SO may affect BO and therefore a need for some method of accounting for and propagating changes to BO exists. I find this claim curious. In the same paragraph, it is said that "it is not clear how to reuse these definitions in BO. Assuming that this problem is solved" then the aforementioned problem may occur. It seems to me that this "assuming that this problem is solved" consists of a knowledgeable person creating BO based on SO. This person would have then created BO based on assumptions of SO. If SO were to change, propagations or effects of these changes occurring in BO may be able to keep BO consistent with SO, but it may not keep the ontology in the mind of the BO-creator or service provider B consistent with those changes. In other words, the bike-store owner still thinks in terms of the old BO and not the BO that results from changes in SO. This seems like it could be confusing. Would a deprecation approach be more suitable? Then the owner/creator/controller of SO could simply deprecate parts of the SO ontology instead of removing parts of it. This would allow for backward-compatability. (This is similar to deprecation of methods in Java.) Is this a viable approach to solving this problem? |
Jesse Weaver |
| Shangguan Managing Ontologies Joshua Shinavier 1 |
The authors mention that Ontolingua supports cyclical ontology inclusion. Is this true of KAON? Does the mechanism of change filtering make it possible to propagate changes through an ontology cycle and its dependent ontologies (that is, without runaway recursion)? |
Joshua Shinavier |
