pbuttigieg
commented
4 years ago I think any "planned P" where P is in scope for ENVO should follow a DP template and always have a logical definition.
Possibly, but I think this branch will organically / sporadically grow, with definitions that won't fit in patterns all the time (some have nuances). We can patternise if patterns are present.
I think it's always good to have clear starts and ends for processes. The above definition and axioms encompasses processes that have deforestation as a part, but may range much further upstream or downstream. For example, many wars would fit this definition.
Sure, but not always feasible or practical. We can always refine the bounds when we have the classes to do so. But adding the various planning processes, consultations, etc (and capturing where different groups believe the begin/end points to be) is a tall order.
I would advocate for something like:
A deforestation that is planned and executed by humans
EquivalentTo: deforestation and realizes some concretizes some plan specification
Ah, so you would constrain this to only the physical deforestation activity itself. That's a good simplification that would allow other classes to capture the non-tree-chopping up/downstream processes, rather than trying to bundle them in.
HOWEVER, I would consider whether we truly need to precompose here. I would much rather provide a more minimal set of classes and have people indicate the degree of planning or other attributes when they use the class.
So would I, but the main user community is not ready to post-compose yet (as you've seen many times). Some are getting there, but there's still a way to go in terms of training.
precomposition makes an ontology harder to maintain and harder for users to use.
Yes, harder to maintain, but at least usable for those that can't postcompose. I think this will be the case until the ontology community reaches out to DMs and curators with a training campaign.
Another oddity here is that the class 'deforestation' has a text definition The removal of forest and undergrowth to, for example, increase the surface of arable land or to use the timber for construction or industrial purposes.
Source: https://www.eionet.europa.eu/gemet/concept?cp=2036 (aside: no longer resolves)
Hmm, GEMET IRIs have changed their form: https://www.eionet.europa.eu/gemet/en/concept/2036
I'll do a bulk replace here as we have a number of other GEMET links. (done in https://github.com/EnvironmentOntology/envo/pull/961/commits/cd82e48c8864c509109a59cdf743575bd1a68f84)
The definition and examples sound indistinguishable from planned deforestation to me, although I think the OBI definition is hard to understand. E.g is slash and burn agriculture a planned process according to the OBI definition?
Hmm, right - I think the GEMET term maps to the planned deforestation class. We'll include the deforestation one as a more inclusive term.
cmungall
I think any "planned P" where P is in scope for ENVO should follow a DP template and always have a logical definition.
Current 'planned deforestation' is:
def A planned process during which humans convert forests into one or more alternative ecosystem types.
SubClassOf:
I think it's always good to have clear starts and ends for processes. The above definition and axioms encompasses processes that have deforestation as a part, but may range much further upstream or downstream. For example, many wars would fit this definition.
I would advocate for something like:
A deforestation that is planned and executed by humans
(I don't love the complexity of the differentia, but this comes from OBI http://purl.obolibrary.org/obo/OBI_0000011)
This makes it a subclass of the existing deforestation class. We have a logical definition, so we get required automatic inferences (subclass of OBI planned process).
This pattern can be applied to any planned variant of an existing environmental process
HOWEVER, I would consider whether we truly need to precompose here. I would much rather provide a more minimal set of classes and have people indicate the degree of planning or other attributes when they use the class. precomposition makes an ontology harder to maintain and harder for users to use.
Another oddity here is that the class 'deforestation' has a text definition The removal of forest and undergrowth to, for example, increase the surface of arable land or to use the timber for construction or industrial purposes.
Source: https://www.eionet.europa.eu/gemet/concept?cp=2036 (aside: no longer resolves)
The definition and examples sound indistinguishable from planned deforestation to me, although I think the OBI definition is hard to understand. E.g is slash and burn agriculture a planned process according to the OBI definition?