Duplicate "Harvard" Styles

The Zotero CSL style file repository contains "Harvard" styles that seem to be duplicates:

Both harvard-anglia-ruskin-university.csl and harvard1.csl are based on http://libweb.anglia.ac.uk/referencing/harvard.htm,

and both harvard-institut-fur-praxisforschung-de.csl and harvard7de.csl are based on
http://www.institut-praxisforschung.com/app/download/6167762784/Harvard-Zitierweise.pdf?t=1410296308 resp. http://www.institut-praxisforschung.ch/Portals/0/Jonas/Harvard-Zitierweise.pdf (the second link is dead, redirecting to http://www.institut-praxisforschung.com/, so this is/was clearly the same guide).

My suggestion is to keep only one style from each pair - if they differ, after checking which one works better -, to continue using the more specific style names - harvard-anglia-ruskin-university.csl and harvard-institut-fur-praxisforschung-de.csl - and to discontinue using the less specific names harvard1.csl and harvard7de.csl.

Bonus: Getting rid of styles with overly generic names such as harvard1.csl and harvard7de.csl would reinforce the important message that there is no such thing as "the" Harvard style, removing a very common source of confusion.
  • FWIW, there is a long and complicated discussion about harvard7de.csl at https://github.com/citation-style-language/styles/issues/74 .
  • I'm generally happy to consolidate Anglia Ruskin and Harvard 1. One problem with this is that Harvard 1 is still a default style in many reference managers (a problem going back to 2008 or so) and so this might be a little tricky, and we'll need to check with other products that this is OK, but as a general preference I think we should still do it.
  • @Rintze: thank you for the link which adds some background the style files themselves just don't provide. Unfortunately, this discussion did not come to a clear conclusion, or at least no action was taken. What both the discussion and the actual behaviour of harvard-institut-fur-praxisforschung-de.csl and harvard7de.csl show, however, is that harvard-institut-fur-praxisforschung-de.csl closely follows Harvard-Zitierweise.pdf, whereas harvard7de.csl does so for the bibliography part only, but modifies in-text citations by adding a comma and "S.", e.g., "(Doe, 2014: S. 12)".

    Since this particular format does not seem to be set down in any single style guide, I would suggest treating harvard7de.csl as a derivative of harvard-institut-fur-praxisforschung-de.csl, and naming it harvard-institut-fur-praxisforschung-modified-de.csl, or harvard-institut-fur-praxisforschung-in-text-citations-with-comma-and-page-de.csl, or something similar. Since the differences are limited to the format of in-text citations, I would also suggest keeping harvard-institut-fur-praxisforschung-de.csl as an independent style and turning harvard-institut-fur-praxisforschung-in-text-citations-with-comma-and-page-de.csl into a dependent one.
  • @adamsmith: That sounds encouraging. On what forum would this need to be announced, or discussed?

    And since this might take some time: What do you think would need to be kept in the meantime? Would it work for legacy programs if just the filename harvard1.csl was kept, but the content was just a copy, or rather, an unmodified dependent style of harvard-anglia-ruskin-university.csl?
  • Would it work for legacy programs if just the filename harvard1.csl was kept, but the content was just a copy, or rather, an unmodified dependent style of harvard-anglia-ruskin-university.csl?
    on a technical level that would work without a hitch. My concern is that we'd be changing a _very_ popular style for Mendeley users essentially under their feet with no easy way to go back. I have no idea if that'd cause a huge outcry or if people using "Harvard1" would even notice.

    The place to bring this up would be xbiblio. I believe you're on that list? In that case you can just describe that proposal (and include the question about removing Harvard1 entirely) and see if there are any objections. Otherwise, I'll raise it when I get a chance, but might take a little while.
  • Just as some background, Carles Pina of Mendeley shared these statistics last year about CSL style usage (I think through their word processor plugins):

    36.8% http://www.zotero.org/styles/apa
    8.2% http://www.zotero.org/styles/ieee
    7.4% http://www.zotero.org/styles/harvard1
    5.1% http://www.zotero.org/styles/nature
    4.0% http://www.zotero.org/styles/vancouver
    2.1% http://www.zotero.org/styles/chicago-author-date
    2.1% http://www.zotero.org/styles/american-medical-association
    1.0% http://www.zotero.org/styles/associacao-brasileira-de-normas-tecnicas
    ...

    (apa is the default)

    So as adamsmith says, harvard1 is extremely widely used, even though we rather would get rid of it.
  • edited September 13, 2015
    Just posted on xbiblio.

    As to harvard-anglia-ruskin-university.csl and harvard1.csl, shouldn't we turn harvard1.csl into an unmodified dependent style of harvard-anglia-ruskin-university.csl, regardless of the outcome?

    Both rely on one single “documentation”, http://libweb.anglia.ac.uk/referencing/harvard.htm, so if the styles differ, they cannot both be correct anyway.

    As to harvard-institut-fur-praxisforschung-de.csl and harvard7de.csl, again both rely on one single source from the “Institut für Praxisforschung”. They do differ slightly (and I was a bit hasty in my xbiblio post here, but then again getting rid of harvard1 is the main goal), but it’s clear that harvard7de is based on “Institut für Praxisforschung”, too.

    Note that http://www.institut-praxisforschung.com/app/download/6167762784/Harvard-Zitierweise.pdf often allows several alternatives, e.g., (Schmidt 2004: 102) / (Schmidt 2004: S. 102) / (Schmidt 2004, S. 102) / (Schmidt 2004, 102); the first being the authors’ recommendation. Since using “S.” (harvard7de.csl) vs. not using it (harvard-institut-fur-praxisforschung-de.csl) is the main difference, this shows rather clearly that harvard7de.csl is just one legitimate variant of harvard-institut-fur-praxisforschung-de.csl.

    Hence I’d propose, again, to rename harvard7de.csl, preferably to harvard-institut-fur-praxisforschung-de-with-s.csl, or harvard-institut-fur-praxisforschung-de-with-seite.csl; title: “Harvard - Institut für Praxisforschung (Bahr & Frackmann) (German, with "S.")”.

    (Precedent, e.g., “American Chemical Society (with titles, no "et al.")”.)
  • I guess you saw the post by Carles Pina on xbiblio.

    It seems Mendeley is not using the renamed-styles.json file, which would speak against retiring harvard1.csl altogether in the immediate future.

    On the other hand, he does not seem to see any problem in turning harvard1.csl into a dependent style of harvard-anglia-ruskin-university.csl.

    So I guess we should do two things pretty soon:

    - Turn harvard1.csl into a dependent style of harvard-anglia-ruskin-university.csl.
    - Rename harvard7de.csl to harvard-institut-fur-praxisforschung-de-with-seite.csl

    Let me know if/when pull requests would be welcome.

    We could also announce that we would like to remove harvard1.csl at some specified date, e.g., mid-2016, giving Mendeley and others time to start using renamed-styles.json, and to communicate the change to their users.

    Thoughts?
  • edited September 25, 2015
    We can also remove harvard1 from the Zotero styles pages, since we do support renamed-styles.json (though we should confirm that it really does work everywhere we need it).
  • We can also remove harvard1 from the Zotero styles pages, since we do support renamed-styles.json
    That might be an interesting experiment, but I rather fix this in the styles repo. A while back we renamed several of the standard Zotero styles, and I don't think we ever heard a peep about that, so that seemed to work. Not sure if turning a style into a dependent works as well in Zotero.
Sign In or Register to comment.