Opened 10 years ago

Last modified 6 years ago

#550 new enhancement

need to automatically strip leading "the" from serial titles

Reported by: erazlogo Owned by: simon
Priority: major Milestone:
Component: styles Version: 1.5
Keywords: Cc: fbennett

Description

citation styles don't use the leading "the" in serials (see 17.195 in Chicago Manual of Style) so it would be great to automatically strip them from imported articles. right now users end up with "The Chronicle of Higher Education" and "The American Historical Review" in the publication field and have to strip those manually. Can this be done in translators?

Change History (6)

comment:1 Changed 10 years ago by erazlogo

  • Owner changed from simon to dstillman

comment:2 Changed 10 years ago by dstillman

  • Owner changed from dstillman to simon

This is universal for all styles? I'm generally wary of doing anything to the data on import rather than export, but if "The" is never useful for anything, it probably makes sense.

Simon's domain, either way.

comment:3 Changed 10 years ago by erazlogo

i'm pretty sure it's universal for all styles (mla, apa definitely). but if this is easier to do during bibliography/note export, that would work also.

comment:4 Changed 10 years ago by stakats

We would definitely want to do this on export, not import. That way we also avoid messing with localization issues – maybe "the" means something else in another language – since citation formats are likely to be more or less local to a given language.

comment:5 Changed 6 years ago by ajlyon

  • Cc fbennett added
  • Component changed from ingester to styles

Is this something that is described in CSL 1.0 and can citeproc-js handle it? It isn't done currently, and I can't find mention of it in the CSL 1.0 spec (hence CC'ed fbennett).

I reclassified this as a style ticket, since I'm in firm agreement with dstillman and stakats that this ought to be an export/style modification.

comment:6 Changed 6 years ago by fbennett

citeproc-js does this in a couple of places currently -- when attempting to generate the citation-label variable (purely experimental), and when composing sort keys, but it's crude, with no sensitivity to user preferences or language selection.

There is some rough infrastructure in citeproc-js to support user-configurable match lists, for various field abbreviations and document-specific "hereinafter" strings. That could be exploited for this, if it comes to be tied into the UI.

Note: See TracTickets for help on using tickets.