material sample versus information artifact
The Biocollections Ontology (BCO) follows other biomedical ontologies by distinguishing material samples from information artifacts at a high level, asserting that hey are fundamentally different. For example, information can be replicated exactly, but samples cannot.
This diagram illustrates that people go out into the natural world and perform processes on material things (organisms, rock, soil, environment [mixed?]) to produce information artifacts or physical samples. The products shown here are just examples for discussion. They may not be correctly named. The list is not complete. For example, tissue-sample could be considered a kind of genetic-resource, and another type of genetic-resource might be DNA-extract.
The principles in designing our subclass hierarchy are:
- a subclass must have (all?) the properties of its parent class(es);
- a subclass is a subset of its parent class; and
- every branch divides a parent class into two or more mutually exclusive subclasses.
(Please correct or add if those statements are not true or insufficient.)
Assertion: providers should characterize their data as specifically as possible in the primary typing scheme, and use other fields to assist aggregators and end users to determine how to handle the dataset or determine the fitness of records.
Would our "problem" be solved by a single field populated with the appropriate name of a leaf node?
Here is slightly more elaborate illustration showing that digital records are eventually created for the different kinds of material samples, and that these can then be used as occurrence data. The main distinction between material sample records and observation records is the reference to the real-world identifier of the sample.
