information-artifact-ontology / IAO

information artifact ontology
Creative Commons Attribution 4.0 International
77 stars 25 forks source link

Annotation to distinguish defined class from universal #52

Open zhengj2007 opened 9 years ago

zhengj2007 commented 9 years ago

From alanruttenberg@gmail.com on August 21, 2009 05:01:48

Perhaps make the distinctions finer along the scheme that Werner has suggested-

Universal Class ad hoc class defined class extension

Need to understand his distinctions first, though.

Original issue: http://code.google.com/p/information-artifact-ontology/issues/detail?id=52

zhengj2007 commented 9 years ago

From alanruttenberg@gmail.com on August 31, 2009 10:38:03

Added draft definitions of denotator type and three instances: universal, defined class, named class expression, in r215 checkin. Needs review.

zhengj2007 commented 9 years ago

From mcour...@gmail.com on August 31, 2009 11:01:52

Could you give an example of what those would be used for?

zhengj2007 commented 9 years ago

From alanruttenberg@gmail.com on August 31, 2009 11:24:27

An example of a named class expression would be the OBI class "material to be added" which is a short hand for material entity and has role some material to be added role.

An example of a universal would be the quality PATO:0001874: disk shaped

An example of a defined class would be Apple or Orange: UnionOf(Apple Orange)

zhengj2007 commented 9 years ago

From alanruttenberg@gmail.com on August 31, 2009 11:26:51

Another example, discussed in IDO recently, would be the class "pathogen", which is a union of all classed defined as beaingr some "pathogenic disposition towards Y", as pathogenic disposition is a relational disposition.

zhengj2007 commented 9 years ago

From mcour...@gmail.com on August 31, 2009 11:43:23

I meant how will the new annotation property be used, and for which purpose. Editor notes state that for universal classes, we could manage disjoints automatically, and that for named classes, we may which not to display them in tools.

Those haven't been major issues until now, and are managed automatically within OBI which is much larger than IAO. Past experience show we haven't been very good at managing curation status for example, and I am just wondering if the benefits to be gained in this case outweigh the curatorial hassle of adding this property to each class?

zhengj2007 commented 9 years ago

From seannos.style@gmail.com on September 01, 2009 01:01:11

Here is my complete list of various sorts of classes that I distinguish for referent tracking:

Defined Class (DCL): a class whose members are particulars which share an additional property but do not form an extension.2

Properly Defined Class (PDC): a DCL whose members are all instances of some universal which share additional properties which are such that (1) there are other instances of that universal which do not share these properties, and (2) there are particulars which do not instantiate that universal but nevertheless exhibit the properties as well. PDCs come in two flavors.

Specifically Defined Class (SDC): each property shared by the members of the PDC is a relationship that they enjoy with another particular, for example all human beings living in Buffalo, NY;

Generically Defined Class (GDC): at least one of the properties shared by the members of the PDC is a relationship that they enjoy with distinct particulars which are all instances of some universal, for example all human beings with tuberculosis. This class is not a SDC because persons can not share their tuberculosis: each disorder in an individual is numerically distinct from a disorder in another individual although these two disorders may be of exactly the same type.

Ad hoc Class (AHC): a DCL whose members are particulars for which there is not a universal that they all instantiate. Here too we distinguish amongst Generic Ad hoc Classes (GAC) and Specific Ad hoc Classes (SAC) in the same way as for PDCs. An example of a SAC is the collection formed by the clinical signs exhibited by a specific patient with tuberculosis. An equivalent GAC would be the collection of clinical signs exhibited by all tuberculosis patients assigned to the control group of a clinical trial.

Werner