ORCID JSON-LD incorrectly asserts that article DOIs are the "sameAs" journal ISSNs!

[rdmpage]

The JSON-LD returned for an ORCID can contain some clearly incorrect assertions. For example, for https://orcid.org/0000-0002-8497-8908 we have entries that include ISSNs, e.g.

The JSON-LD for this record looks like this:

{
	"@type": "CreativeWork",
	"@id": "https://doi.org/10.1007/s12686-010-9350-y",
	"name": "Thirteen polymorphic microsatellite markers for the European green toad Bufo viridis viridis, a declining amphibian species",
	"identifier": [
		{
			"@type": "PropertyValue",
			"propertyID": "issn",
			"value": "1877-7252"
		},
		{
			"@type": "PropertyValue",
			"propertyID": "doi",
			"value": "10.1007/s12686-010-9350-y"
		},
		{
			"@type": "PropertyValue",
			"propertyID": "source-work-id",
			"value": "cv-prod-id-1913589"
		},
		{
			"@type": "PropertyValue",
			"propertyID": "issn",
			"value": "1877-7260"
		}
	],
	"sameAs": [
		"https://portal.issn.org/resource/ISSN/1877-7252",
		"https://portal.issn.org/resource/ISSN/1877-7260"
	]
}

The sameAs assertions are clearly false. The article with DOI https://doi.org/10.1007/s12686-010-9350-y is not the same thing as the journal with ISSNs 1877-7252 and 1877-7260.

Without looking at the source code I assume there is a blanket rule that any resolvable identifier in addition to a DOI is automatically an identifier for the article. This is clearly incorrect.

[rdmpage]

I note that records where the web page displays the article as having a Part of relationship with on for more ISSNs don't display ISSNs in the JSON-LD, e.g. the record for https://doi.org/10.11646/zootaxa.4759.4.12 for the same ORCID:

[TomDemeranville]

Thanks for the feedback and sorry we missed this.

As you can see, when members add the correct metadata we don't use sameas for the ISSN.

The first example is essentially saying they created a journal issue (which is a thing for guest editors can do for specific issues of an ISSN) or are editor in chief etc. Which is incorrect. The issue here is with the metadata supplied by the ORCID member that wrote the item to the record.

[rdmpage]

@TomDemeranville I understand that the problem is likely to be with the data provider CIÊNCIAVITAE. Hell is other people's data 😉

I dont think this is a case of

The first example is essentially saying they created a journal issue (which is a thing for guest editors can do for specific issues of an ISSN) or are editor in chief etc.

Rarther it looks like an article record that has ISSN information recorded in a way that associates it with the article not the journal. Looking at the record online for this person https://www.cienciavitae.pt/portal/B41A-54F4-007F I can't see any difference with the other articles, presumably the problem lies in whatever file format CIÊNCIAVITAE send to ORCID.

In any event, it would be useful if ORCID had internal rules that caught things like this so that anyone using the RDF didn't end up making obviously wrong assertions (i.e., that an article is the same as a journal).