Hi Dan,
On 12/12/2019 16:30, Stoner, Dan F wrote:
I found some oddities and I am not exactly sure where to go next.
We are noticing the following while processing meta.xml in darwin core archives produced by IPT (and other servers):
Schema validation failed, continuing unvalidated XMLSyntaxError: Element '{http://rs.tdwg.org/dwc/text/%7Dcoreid': This element is not expected. Expected is ( {http://rs.tdwg.org/dwc/text/%7DcoreId )
There's some background to this on this issue: https://github.com/tdwg/dwc/issues/143
The schema itself and the documentation were conflicting, and this was fixed (in mine and Tim's opinion) the wrong way, by changing the schema.
*I've just pushed a commit to fix it the right way,* i.e. reflecting 99% actual usage and leaving the schema as it was for almost a decade.
Although we do accept either, we still see only 31 datasets registered in GBIF with "coreId" rather than "coreid".
It seems like most consumers are not actually validating meta.xml using the schema, and the producers are generating files out of compliance with the schema.
Most of the Darwin Core archives I have manually inspected and tried to validate contain meta.xml with lowercase "i" in coreid despite the Standard indicating capital "I" in coreId.
I poked at the GBIF Darwin Core Validator 3 code repo and found this:
schema.meta=https://raw.githubusercontent.com/tdwg/dwc/master/standard/documents/text/td...
The first link leads to 404, the second leads to an xsd that contains the proper coreId. So maybe the Validator is not being "strict" about validation against the schema?
I suspect it has been running for so long that, when the validator process was originally started, both URLs were valid, and had coreid or one of each.
Cheers,
Matt