tdwg / esp

Earth Sciences and Paleobiology Interest Group
13 stars 10 forks source link

How can we indicate an intentionally empty value vs a Null value? #12

Open dennereed opened 7 years ago

dennereed commented 7 years ago

Some Darwin Core fields may not apply or may be intentionally blank. What is best practice to distinguish blank entries from Null entries, which are not blank, nor filled, but unknown.

DimEvil commented 7 years ago

I would suggest: If the Darwin Core fields do not apply, don't publish them. If no information is withheld, don't publish the term. If the same information is withheld over the dataset, use a fixed term: , [informationWithheld] = 'Length and weight measurements available upon request.'

if the information is subject to a specific record, only provide this information in the specific record, the other field are left empty.

dennereed commented 7 years ago

This is a general Darwin Core issue, but is particularly salient for paleo because there is often greater uncertainty for paleo specimens. The Darwin Core QA thread has a similar question regarding uncertainty for sex. It's identified as an issue of structured vocabulary. Having terms to express unknown and uncertain. Creating a FAQ on this topic.

garymotz commented 5 years ago

In the Interest Group meeting at Biodiversity_next today (the combined TDWG:Paleo Interest Group with the ABCD:EFG and iDigBio Paleo Working Group), it was discussed that there was a use case in which intentionally blank values in dwc:basisOfRecord, indicate "recent" specimens in a paleo/earth science collections context. It seems, therefore, they're a collection of typically non-recent collections objects in which some "recent" objects are present and the absence of the dwc:basisOfRecord = "fossilSpecimen" in their fossil collection indicated a "misfit" specimen not deemed appropriate for deaccession.

NB - I'm attempting to document this use case here in an attempt to demonstrate to others what this unique situation might contribute to, or confound, the "proper" use of blank entries in certain necessary Darwin Core terms.