Upgrade StorageJournal abbreviation – bug report and questions
Here’s a bug report and a number of questions on Zotero’s journal abbreviation mechanism:
(1) Bug report
Setup: Zotero/LO, cell.csl
Item 1:
Publication = "Deutsches Archiv für klinische Medicin"
Language = "de-DE"
(a) with "Automatically abbreviate journal titles" in "Document preferences" unchecked
Expected output: "Deutsches Archiv für klinische Medicin"
Actual output: "Deutsches Archiv für klinische Medicin" (i.e., as expected)
(b) with "Automatically abbreviate journal titles" in "Document preferences" checked
Expected output: "Dtsch. Arch. klin. Med."
Actual output: "Dtsch. Arch. Für Klin. Med." (i.e., "klin." incorrectly capitalised, and the preposition "für" not removed)
Note that the capitalisation of non-English items should never, ever be modified – and this includes abbreviation mechanisms, too.
(2) Questions
Zotero’s journal abbreviation mechanism seems to use (a local copy of) https://raw.githubusercontent.com/zotero/zotero/master/resource/schema/abbreviations.json – correct? Where/under which name is the local copy stored?
This file contains two sections, "container-title" and "container-title-word". While I’m not sure where exactly the "container-title" items come from, the "container-title-word" section is obviously derived from the ISO 4 “List of Title Word Abbreviations (LTWA)” available at http://www.issn.org/services/online-services/access-to-the-ltwa/. (Note that a pdf of the complete list is made available on that page, but it seems not all words with accented characters can be extracted reliably from that pdf, at least not via pdftotext. However, the complete list can also be downloaded from the html pages
http://www.issn.org/services/online-services/access-to-the-ltwa/?lettre=&numpage=1&numpagemod=#lettres to http://www.issn.org/services/online-services/access-to-the-ltwa/?lettre=&numpage=2802&numpagemod=#lettres.)
The last 126 lines of "container-title-word", however do not seem to be from LTWA. They appear to be prepositions in various languages, including a "fur" which I would guess is a corrupted form of the German preposition "für" – so I would suggest replacing "fur" by "für".
There are also numerous "fur"s in what all seem to be German titles in the "container-title" section, so I guess all of these, and possibly others, will need to be fixed, and capitalisation, too, will have to be put right.
(1) Bug report
Setup: Zotero/LO, cell.csl
Item 1:
Publication = "Deutsches Archiv für klinische Medicin"
Language = "de-DE"
(a) with "Automatically abbreviate journal titles" in "Document preferences" unchecked
Expected output: "Deutsches Archiv für klinische Medicin"
Actual output: "Deutsches Archiv für klinische Medicin" (i.e., as expected)
(b) with "Automatically abbreviate journal titles" in "Document preferences" checked
Expected output: "Dtsch. Arch. klin. Med."
Actual output: "Dtsch. Arch. Für Klin. Med." (i.e., "klin." incorrectly capitalised, and the preposition "für" not removed)
Note that the capitalisation of non-English items should never, ever be modified – and this includes abbreviation mechanisms, too.
(2) Questions
Zotero’s journal abbreviation mechanism seems to use (a local copy of) https://raw.githubusercontent.com/zotero/zotero/master/resource/schema/abbreviations.json – correct? Where/under which name is the local copy stored?
This file contains two sections, "container-title" and "container-title-word". While I’m not sure where exactly the "container-title" items come from, the "container-title-word" section is obviously derived from the ISO 4 “List of Title Word Abbreviations (LTWA)” available at http://www.issn.org/services/online-services/access-to-the-ltwa/. (Note that a pdf of the complete list is made available on that page, but it seems not all words with accented characters can be extracted reliably from that pdf, at least not via pdftotext. However, the complete list can also be downloaded from the html pages
http://www.issn.org/services/online-services/access-to-the-ltwa/?lettre=&numpage=1&numpagemod=#lettres to http://www.issn.org/services/online-services/access-to-the-ltwa/?lettre=&numpage=2802&numpagemod=#lettres.)
The last 126 lines of "container-title-word", however do not seem to be from LTWA. They appear to be prepositions in various languages, including a "fur" which I would guess is a corrupted form of the German preposition "für" – so I would suggest replacing "fur" by "für".
There are also numerous "fur"s in what all seem to be German titles in the "container-title" section, so I guess all of these, and possibly others, will need to be fixed, and capitalisation, too, will have to be put right.
It appears that Zotero uses a hybrid of NLM-Medline standard abbreviations and when those aren't available goes to the ISO-4 LTWA standard. I should point out that there are words to be abbreviated that are not (yet) included in the LTWA lists. Those words are either used without abbreviation or, if abbreviated, altered to follow a standard pattern for abbreviation of terms.
Articles and prepositions (with _very_ few exceptions) should never be included in journal abbreviations with the NLM / LTWA rules.
Under NLM standards each part of the abbreviation first chaaracter is in upper case and diacritical marks are removed. The LTWA standard usually includes diacritics and follows the case standard of the journal language.
To complicate matters, both the old AACR-2 and the current RDA standard cataloguing rules encourage using upper case _only_ for the very first character of the first abbreviated word and the rest all lower-case.
Things can be less straightforward:
Both BIOSIS and CASSI abbreviations are based upon LTWA but often with modifications.
Engineering, law, mathematics, and other disciplines and their journals have still more listings of abbreviation standards.
To further complicate the issue, many journals send non-standard abbreviations of their titles with their metadata. Some of these can be quite strange.
This does not consider the many journals with titles that do not use the Roman alphabet where a title is transliterated (by entities other than the publisher) into a Roman alphabet title.
I welcome the comments of the librarians who participate in this forum.