#777 modeling data item

[jzheng]

During OBI core term review in 2013, we had extensive discussion of data item and changed the modelling of measurement data using value specifications. Details please see:
https://svn.code.sf.net/p/obi/code/trunk/src/examples/development/data-prototype.pdf

However, this model is not fully implemented in OBI and applied to one term 'predicated mass value'. There are inconsistent data item model implemented in OBI (different from data item representation in IAO). For example,
http://purl.obolibrary.org/obo/OBI_0001929 mass value specification
http://purl.obolibrary.org/obo/IAO_0000414 mass measurement datum

In addition, there are two topics we have not covered in previous data item discussion.

1. how to logically distinguish measurement datum, predicted datum, simulation datum, calculated datum, etc.

2. data scales

For example, age data can be 1, 2, 10, ... year (s), or child, young, and old. IAO defined ‘categorical measurement datum’ and ‘scalar measurement datum’. So, for 'age datum', we may need to define, age datum, scalar age datum and categorical age datum. It will have multiple asserted inheritance issues if we define in this way. Need to define categorical measurement datum and scalar measurement datum using equivalent axioms, such as,
scalar datum = data item and has_value_specification only 'scalar value specification'
categorical datum = data item and has_value_specification only 'categorical value specification'

Therefore, I’d like to discuss the ‘data item’ again.

Mathias and I discussed about simulation and prediction data after OBI data item and documented on the google doc:
https://docs.google.com/document/d/1VMDp1WsDiC6n54mcCRrMehnalo7_P_MrXqBimfyPw2w/edit?usp=sharing

It listed the issues we thought about and proposed solutions. It may help on our data item discussion.