ANZSoilData / def-au-domain

A domain-model for soils, in the form of a set of classes and properties that define the entities ('features' in GIS-speak) that comprise the important concepts in soils.
Creative Commons Zero v1.0 Universal
4 stars 0 forks source link

Cardinality of profile parameters #37

Open broughd-qld opened 1 year ago

broughd-qld commented 1 year ago

Is there a need to specify a cardinality (i.e mandatory requirement) for some of the properties? Example and discussion below is for coarse fragments, but equally applies to many property types.

The old SITES schema had a mandatory constraint for certain parameters at the database level and this has been continued into this new soil information model. As an example refer to the requirement to have a coarse fragment abundance.

https://github.com/ANZSoilData/def-au-domain/blob/ea9af3883b40daed69a976f32454fd2675a1ff21/rdf/properties.ttl#L155-L159

By setting this cardinality constraint there is a need to include an additional term in the vocabulary for coarse fragment abundance for a 'Not Known' value. There is a lot of data where the abundance is not known, but other properties about coarse fragments are. Should that data be discounted because of a lack of abundance information?

I suggest there is a relaxation to the mandatory nature of the requirement and an additional constraint of at-least one of the coarse fragment properties.

dr-shorthair commented 1 year ago

I'm just in process of preparing an alternative schema which uses references to entries in RVA, in place of the codes. Not sure which is preferred.

But if vocab references are used, then it will be easy to add 'no known' to all the vocabs. OTOH if we stick with codes, then there would need to be a 'not known' code added to all the enumerations. Given the code lists are very heterogeneous, we'd probably have to go to something like "00" for "not known" since I don't believe that code is used anywhere.

Don't quite understand your last sentence.

broughd-qld commented 1 year ago

RE the last sentence - If you relax the restriction on the abundance, you could an an option that for a Coarse Frag record to be valid it needs to include one of the parameters in Lithology, distribution, shape, size, strength and abundance.

dr-shorthair commented 1 year ago

Ah - OK. We did discuss this. I'll see if I can tweak it that way.