WoRMS Polychaeta (id 1090): test report

yroskov commented 3 years ago

WoRMS Polychaeta, id 1090 on prod https://data.catalogueoflife.org/dataset/1090/meta

[x] Sector: class Polychaeta (was not broken)
[x] Metadata: as available for 2021-04-01. "Read, G. Fauchald, K. (Ed.)" was placed in Authors field. Corrected in clearinghouse.
[x] Imported: 12664 spp vs 12270 in ac19

yroskov commented 3 years ago

ISSUES (selected only) 2021-04-21

[ ] Unparsable Year, 40 - @mdoering, what might be wrong with years in these names? They look fine for me. https://data.catalogueoflife.org/catalogue/3/dataset/1090/workbench?facet=rank&facet=issue&facet=status&facet=nomStatus&facet=nameType&facet=field&facet=authorship&facet=authorshipYear&facet=extinct&facet=environment&facet=origin&issue=unparsable%20year&limit=100&offset=0

yroskov commented 3 years ago

TASKS 2021-04-21

Resolved:

yroskov commented 3 years ago

Synced 2021-04-21

mdoering commented 3 years ago

The verbatim browser can also filter by issues, but there is no UI filter for that yet. If you add it manually to the URL you can see the verbatim records with that problem:

https://data.catalogueoflife.org/catalogue/3/dataset/1090/verbatim?issue=unparsable%20year

Screenshot 2021-04-21 at 23 39 32

You can see immediately that the publishedInYear field is the problem. There are original dates in there which is fine, but the parser seems to have problems deriving a year from that. What do the rectangular brackets mean again?

There also seems to be another semantics/mapping problem here. Nereimyra mutilata is a recombination with the authorship (Treadwell, 1901). The publication year and reference should never reflect the basionym but the publication that created that specific combination. If that is unknown referenceID and all publishedIn fields should stay empty.

ColDP specifies:

a Reference indicating the original publication of the name in its given combination, not the basionym.

bart-v commented 3 years ago

ICZN art. 22A.2.3.

if wishing to cite both the actual and the imprint dates, should first cite the actual date (cited as above), followed by the imprint date for information and enclosed in parentheses or other brackets and quotation marks; (...)

https://code.iczn.org/date-of-publication/article-22-citation-of-date/?frame=1

bart-v commented 3 years ago

Correct about Nereimyra mutilata and publishedIn This is now fixed for the next export

yroskov commented 3 years ago

What do the rectangular brackets mean again?

It means, that taxonomic expert has doubts about the year in the authorstring or in the reference; a year should be confirmed in further research. Unconfirmed year might be indicated with question mark or put in rectangular brackets.

yroskov commented 3 years ago

The publication year and reference should never reflect the basionym but the publication that created that specific combination.

It is true for botanical names, but incorrect for zoological.

bart-v commented 3 years ago

It is true for botanical names, but incorrect for zoological. Good point.

Then then the question is do we keep the current definition of PublisheIn or not?

a Reference indicating the original publication of the name in its given combination, not the basionym.

This is quite clear...

yroskov commented 3 years ago

I meant in my sentence, that zoological names do keep an author and year of original combination (= basionym in botany) only. A year of the publication, where consecvent combination was published, usually, are not present in the authorstring of that combination. So, a year in the bracketed authorstring, usually, is different with a year in the reference.

mdoering commented 3 years ago

Yes, that's correct. But publishedInYear and the other publishedIn and referenceID fields are explicitly defined not be the basionym/protonym, otherwise we interpret things wrongly. I know it is not common to track or even publish a new combination in zoology (with some exceptions in some groups, it snot the same practise everywhere). But in those cases KEEP THE FIELD EMPTY. Don't populate it with wrong years. We code against the expectation that this is the original publication of the combination.

mdoering commented 3 years ago

When the referenceID (or publishedInYear if the entire reference is not known) is correct for combinations you can e.g. organise them chronologically like Plazi does:

mdoering commented 3 years ago

Note also that ColDP says about publishedInYear:

The effective year the name was published, given as a 4 digit integer. It is the year that is nomenclaturally relevant for the given combination. In most cases this will be the same as the publication year given in the linked reference record via referenceID. But in some cases this might be different.

So ideally imprint years and uncertainty markers should not be found in the field. They can exist in the authorship, but this field is in particular meant for understanding the effective year of publication.

I will make sure the interpreter understands also imprint years and uncertainties, but in a perfect world these should not exist here.

yroskov commented 2 years ago

ver 2021-11-01

TASKS - there are changes