The file term_versions-old.csv is the "normative document" modified according to the various commits documented in GitHub. term_versions.csv was generated by the build script without the term-localname column. The generated_normative_document.csv was generated by the build script with the term-localname column added.
The comment "Access Rights may include information regarding access or restrictions based on privacy, security, or other policies." is found in the comments for dcterms:accessRights, not as part of the definition.
The DCMI History page http://dublincore.org/usage/terms/history/ uses a version model that differs from what we are using in TDWG. They use the term '"issued" for what we call "current terms", whereas we only use "issued" for term versions. Thus their issued date is what we call the created date and their modified date for the current term is what we call the issued date of the version.
These dates have been changed to match the TDWG model. Also, version 003 of dcterms:modified has "Date Modified" as the label, rather than "Modified"
dcterms:rights was originally in Darwin Core, but was replaced by dcterms:license in Executive Decision http://rs.tdwg.org/dwc/terms/history/decisions/#Decision-2014-11-06_17 . It has been added as part of the historical record. It is listed as the "replaces" value for dcterms:license term and needs to be there.
In the previous term_versions.csv file, the value in the replaces column indicated what DCMI said the term replaced. However, in the complete history table, we use this column to indicate what this term version replaces with respect to previous TDWG use. For example, in an Executive Decision, dcterms:license was to be used in lieu of dcterms:rights. I've changed the table to reflect this replacement, rather than indicating how http://dublincore.org/usage/terms/history/#rightsHolder-002
was caused by DCMI to replace http://dublincore.org/usage/terms/history/#rightsHolder-001 . There are several other examples of replacements caused by TDWG actions and these have replaced the record of DCMI actions. If there was no TDWG-sanctioned replacement, the column was changed to empthy.
Note: I did not change the obsolete terms, which have always had camelCase labels. Also, the IRI terms that have non-IRI analogs have " (IRI)" appended to their labels to distinguish them.
Move dcterms:type below the recommended terms and change status to deprecated. Change its notes to indicate that dc:type should be used for strings and that the RDF guide recommends rdf:type instead of dcterms:type when providing an IRI value. Replace dcterms:type in the record-level terms with dc:type and carry over the examples that were previously in dcterms:type to it. NOTE: the terms below the recommended terms are not in alphabetical order still.
In the record level terms section, replace dcterms:language with dc:language. Use the former dcterms: examples for string values with dc:language. Move dcterms:language to the UseWithIRI section and change its recommended value to an IRI from the LOC ISO 639-2 scheme (to maintain consistency with Audubon Core usage. NOTE: the value of organized_in for dc:language is now different from the other DCMI terms in the record-level section. Will that break the QRG build script?
So they are mixed with the order recommended terms, in Darwin Core order. Ran python build.py after this and there are no changes in the quick ref guide
See #262:
- Remove 160 terms from 2018-09-06 that are unchanged duplicates of the 2017-10-06 version
- Reinstate "recommended" status for the 160 2017-10-06 terms 45 + 47 + 36 + 1 + 31
Not that the reinstated recommended terms are NOT YET SORTED to the top
Steve Baskauf
John Wieczorek
peterdesmet
Matthew Blissett
Peter Desmet
Andrew Punnett