EDIT

Questions:

• how to map reference type article? "article" or "article journal"?
• how to handle nomencl. referneces in the reference table as they usually do not have full titles and authorship differs => there are 3 new fields now (see above).
• how to handle details/microReferences, e.g. for nom. refs.

• ref.volume, .issue and .edition that are not numbers (and .issue in general as we don't split this)

• is distribution status "domesticated" relevant for plants? (http://api.checklistbank.org/vocab/distributionstatus)

• should doubtfully native be included in "native" and should "unknown endemism" be include in native?

• what to use for the nomenclatural code? The name or the acronym? (http://api.checklistbank.org/vocab/nomCode)

WGB:

I looked at the CoLDP documentation and I cc to Markus (who is on vacation I believe).

Reference.type: I think our articles are always journal articles and should be mapped to this term (in Zotero there is no generic “article”). If there would be a lone newspaper or magazine article somewhere in our data, it wouldn’t matter, because the format is practically the same.

nomenclatural references:

I don’t think that differing authorship is a problem, because the nomenclatural author is handled by the Name table: Name.authorship. Here the standard abbreviated author should be entered for botanical names.
The reference detail is also handled there: Name.publishedInPage.
Having the Name.publishedInYear here as well refers more to zoological names, but should be filled for botany as well, because it can be useful to distinguish isonyms etc.

Indeed, the rest of the citation has to be handled in the Reference table (this is also true for the simplified NameUsage table).
I see no way here to have bibliographical and nomenclatural refereces for botany represented in a single record – the standard abbreviations cannot be entered in parallel to the full title. So separate records have to be created. (No title abbreviation in BibTex neither, I believe).
The relationship between nomenclatural record and bibliographic record that we support in the CDM is thus lost.
I see no tricks to get around that - since there is no recursive relationship for the Reference table (“containerID”), this cannot be solved in the current CoLDP model. However, avoiding recursive relationships in an exchange model definitely helps.

@Markus: This could be overcome by a TEXT field in Name for nomenclaturalReference or in Referendce for nomenclaturalTitleAbbreviation. The latter would be more normalised, the former would have the advantage to have all name data in the Name table.

MD:

just quick from Cologne: ColDP offers 3 ways to share structured references, CSV, BibTex or CSL-JSON.
CSL-JSON is extensible and has many fields that can be used, including "short" titles like title-short to hold an abbreviated version:
https://github.com/citation-style-language/schema/blob/master/schemas/input/csl-data.json
It is supported by most of the large reference information systems like CrossRef, DataCite, etc and you can resolve any DOI to it easily.

Providing JSON is rather difficult for most people though, so I have created an issue to extend the CSV format to also be able to do the same, also for the author and journal: https://github.com/CatalogueOfLife/coldp/issues/74
If you want use that already. It will take some time to be properly supported, but should happen some time this year for sure. It would be a shame to have 2 references for the same thing.

As Walter said the Name.publishedInYear is useful for botanical and zoological names. Especially for botany it allows to compare and sort names on their year of publication, which is otherwise not existing in the authorship. And not in all cases a reference record exists.