samleeflang
commented
1 month ago Hi David, Thanks for creating this ticket. I understand what you mean, but there are some reasons why we did follow this approach.
The first is that we followed the setup used by LatimerCore on their terms page https://ltc.tdwg.org/, setup by Ben Norton. This has been ratified, and we felt it has a clear and reusable structure. If you look at their terms pages it has the same issue as you describe for DiSSCo, for example there is:
- https://ltc.tdwg.org/terms/index.html#ContactDetail_hasIdentifier
- https://ltc.tdwg.org/terms/index.html#EcologicalContext_hasIdentifier
- https://ltc.tdwg.org/terms/index.html#ChronometricAge_hasIdentifier
The term
ltc:hasIdentifieris used in several classes and is therefore replicated several times on the terms pages. Of course, this doesn't mean that this might not be want to change it, but it does mean that at least the structure should be recognisable and has been accepted by TDWG.
Another argument to replicate terms is to easily give the user a complete overview of all terms in an object without having to move back and forward between different pages or part of a page. We wanted to keep things simple and together, grouped around the class it is a part of. If we would deduplicate the structure we would need to create additional pages for the nested objects (Agent, Identifier, etc...) resulting in even more pages. Then, if you want a complete overview of the Digital Specimen you would have to flip between pages, check the “Belongs to Class(es)” and construct the object. As these are pages for human use (schemas.dissco.tech is for machines) we felt we should keep it as simple as possible.
However, that does not mean that this isn't a valid point, and we would have to see how this works out when we create the openDS terms. Let's see this as a first version, on which we iterate and improve.
Kind regards, Sam
Suggestion
For each FDO, there are 3 resource pages, the terms page, the quick reference guide and the core resources. the terms page and the guide are very close to each other in terms of the information they present. This structure is taken from the DwC documentation, it differs however in certain ways, that are a bit of a disadvantage.
In the guide, the terms are structured a bit more user friendly. All of the properties are sorted by their class, even if this means that reused properties occur multiple times on that page. On the terms page, the terms show a bit more information.
When the URI of a term in DwC is resolved, it is redirected to the terms page, which is sorted alphabetically. DwC doesn't have any reuses of properties between classes. The ones that are, are sorted in their own group called "Record level". Which is why they don't encounter a problem that exists for openDS.
The openDS terms don't resolve yet and I am not sure, what the planned strategy is for when they resolve, but at the moment they could not link to the correct definition, as the entries on the terms page are still dependent on the class where they are used. Compare for example MaterialEntity:organisationName and DigitalSpecimen:organisationName, both having the same URI, but different documentation here, in regards to the "Belongs to Class" segment.
We feel that the terms pages need to be restructured, so that each term only occurs once. The terms could be sorted alphabetically and if reused by different classes, it should list them as "Belongs to Class(es):" . The anchor links at the section Index By Term Name should only include the term name and not the class, just as the DwC documentation does.