In filing away messages, I noticed that a response was (I think) requested from me here. I am going to dodge answering the question for two reasons: first, it wasn't my preference to address the issue of heterogeneous composite samples through defining them as Individuals and second, I think Rich's answer was as good as any I could give. However, in typical evil teacher fashion, I'll answer the question with a question. In the existing system where Individuals are denormalized out of the picture and Identifications are applied directly to Occurrences rather than Individuals, did anyone ask the same questions about Occurrences? If one is recording an Occurrence of a lichen, human, or coral is it required that the Identification make the distinctions you are asking for here? I think we have been satisfied with an level of sloppiness in semantics in that case that is equal to what we may have to put up with if we allow Individuals to exist as a class.
Steve
Bob Morris wrote:
When you guys finally agree (or not) here are some questions as a developer that I would ask you each. They are not that different from the issue I raised timidly about tying the semantics of "Individual" to distinguishing the "origin" of an aggregation. As always, pardon any misuse of biological terms.
- Neglecting-if one can--issues about colonial organisms--is a lichen
one individual or two? 2. In a way consistent with your answer to 1, counting the human obligate symbionts such as gut bacteria, is a human one individual or many? 3. If you feel no compulsion to be consistent in answering 1 and 2, will the addition of class Individual to DwC require further properties to determine which of your inconsistent uses is in play, in order that semantic integration about data on Individuals not become logically inconsistent? 4. In the case of lichens or other(???) "compound" organisms whose taxon name is conventionally given by the name of the fungal component, are new DwC terms needed to distinguish whether an Individual of that name is a lichen or a fungus?
(As far as I can tell from a bit of browsing on the web, the current answer to the conundrum of 4 seems to be that the distinction is made mainly in the dataset metadata, declaring in some way that the dataset is of fungi or is of lichens. This doesn't seem very satisfactory in a world where aggregators may isolate data from the datasets. It probably imposes higher-than-record-level provenance requirements on the integration.)
Bob Morris