Open SteffenBrinckmann opened 2 months ago
A few remarks while manually checking the JSON-LD:
sdPublisher
, author
, etc. should ideally be flattened and only referenced via their @id.
author
or the flexible metadata are missing a @type
.variableMeasured
is the best place for the flexible metadata beyond the initial suggestion?SampleDB currently fails to import the .eln file as the file ./temporary_pastaTest/StandardOperatingProcedures/SEM.md
is referenced, however it does not exist. The temporary_pastaTest
directory only contains a file data_hierarchy.json
, nothing else.
Aside from that, there seem to be a lot of additional properties there, which I'm not sure whether they are useful to display to a user, e.g. _attachments.v0.json.digest
to _attachments.v51.json.digest
(and the respetive lengths, revpos, etc). These are also listed for File
objects, not just for Dataset
objects, which isn't valid.
The example directory README should be updated to reflect the new ro-crate-metadata.json
, right?
@nicobrandt I don't think so, but going by schema.org, a Dataset
can contain a PropertyValue
either as variableMeasured or variablesMeasured, aside from it being used as identifier. variablesMeasured
is retired and recommends using variableMeasured
, so unless we want to deviate from the existing properties, variableMeasured
is the most fitting.
A few remarks from my side:
hasPart
property in the second object in the @graph
, should have only the relative paths of the folders. It should be updated in the specification too.{ "@id": "./temporary_pastaTest/Steel/" }, { "@id": "./temporary_pastaTest/SurfaceEvolutionInTribology/" },
are listed in the hasPart
but are not present in the archive.@SteffenBrinckmann, a few remarks:
ORCID
in the @author
node, you should use identifier
:"identifier": "https://orcid.org/0000-0003-3432-2297"
shasum
property? The hashing function name must be mentioned.@nicobrandt @jmanideep @FlorianRhiem @NicolasCARPi Thank you all for the feedback: helpful and I will incorporate it in the next version. We have not agreed an all of the "methods" used here and the idea here is not to fix anything. It is meant as a test area for ideas and see if/how they could work. Trying something in code is - in my view - better than theoretical discussions. Once we finally agree on things, then we can write tests for them and ... . Nevertheless, this is just the output of code and that code is easily changed, no hard feelings if we agree on alternative methods.