Apparent Change in Style Processor

A change has apparently been made in the style processor. I think it used to substitute container-title if the style accessed container-title-short and no value existed. Now it doesn't.

After updating a reference list there were dozens of entries with no journal name. We had to go into each entry in Zotero and make entries in the Journal Abbr field. Can this be changed back to the way it was before?
  • which citation style?
  • edited August 23, 2013
    It's a version of the Pharmacoeconomics style that I customized myself. A few weeks ago it worked, today it doesn't. Only change is that -- on yeah -- I started using zotero standalone.

    Am I crazy? When you specify container-title-short and there's nothing in that field, isn't it supposed to then use container-title? Does the style processor in ZS work differently than ZFF? Has this feature changed in the style processor?

    Anyway, they worked before. Now I had 2-3 dozen entries without journal names in the reference list.
  • edited August 23, 2013
    might have changed in the processor, but if so I'd argue the change is preferable - a variable shouldn't just silently substitute. Also it gives you more flexibility:

    You can still get the substitute behavior by using
    <text variable="container-title" form="short"/>
    which is what all the "official" citation styles use.

    edit: FWIW behavior for title-short is the same.

    edit2: this also seems to be in line with the specs. The substitute behavior is described for form="short", not mentioned for (container-)title-short
  • edited August 23, 2013
    It's pretty jarring when you're about to send revisions to a medical journal and suddenly a slew of references no longer have journal names.
  • BTW, I checked ZFF. Same behavior.
  • not sure what to say. This is relatively rare, but those types of processor changes happen, especially when they bring the processor in line with published specs.
  • I'll make the change to our version. Someone might look at the Pharmacoeconomics style to see if it has the same problem. I didn't build our style from scratch.

  • I checked, it doesn't. There was one style on the repo that used container-title-short, I changed that.
  • The visual editor at makes short work of a change like this.
  • For what it's worth, the "-short" variables were specifically introduced in CSL to have a way to avoid automatic substitution by the unabbreviated form if the abbreviation is not available. So the new behavior is definitely correct, and it's indeed the style files that need to change.
  • Does zotero automatically update styles from the repository?
  • In 4.0, yes.
  • Is there a way to get on a mailing list for changes to the style processor?

    We need a way to keep our citation/reference styles stable. We can do that by rolling our own, but that doesn't insulate us from changes in how they are processed.

    (I know the above sounds selfish, but I've never wanted to upload updates to styles because a change that's good for us might be bad for someone else.)
  • Changes to the citeproc-js CSL processor used by Zotero are logged at
  • And the Zotero changelog says when new versions of citeproc-js are integrated into Zotero.
  • the xbiblio devel list that covers major processor changes is here:
    but updates that just fix mistakes as this one did won't be discussed on the list. The only other place to look would be the citeproc commit log:
    (it's also possible that this behavior was introduced by a change in how the Zotero world plugin handles abbreviated journal titles. In that case it wouldn't have been clearly visible anywhere).
  • I saw this behavior this morning. Our Zotero entry in the publication field was "Value in Health: The Journal of the International Society for Pharmacoeconomics and Outcomes Research". The entry in the Journal Abbr field was "Value Health".

    Zotero produced this entry for the reference: "Value Heal J Int Soc Pharmacoeconomics Outcomes Res". This is NOT the National Library of Medicine abbreviation for this journal. "Value Health" is the abbreviation.

    Where is the abbreviation coming from? The style uses this code:<text variable="container-title" form="short"/>
  • edited September 6, 2013
    I looked further in the reference list and found NUMEROUS non-standard abbreviations of journal names.


    Medical journal submissions MUST conform to NLM journal abbreviations.
  • When will this be fixed?
  • I see from " A way to batch edit journal abbreviations?" that you've changed the behavior of the Word plugin without telling anybody.

    Changing functionality without notifying users is a sure way to move them to Endnote or some other platform.

    1) set the default for the journal name abbreviation checkbox on the Word plugin to be unchecked
    2) email all users when you make a change in functionality
  • edited September 6, 2013
    Zotero will automatically abbreviate "Value in Health" correctly, but it doesn't know how to deal with the subtitle.

    Automatic journal name abbreviations are disabled by default for existing documents. Zotero uses the setting you most recently selected as the default for new documents.

    Nothing about this behavior has changed since the release of Zotero 4.0. We aren't going to email all of our users every time we update Zotero, but all changes are documented in the changelog, which we update before every Zotero release.
  • @rwielage, see (especially "AVOID ALL-CAPS").

    Zotero's automatic title abbreviation should be better documented, though.
  • Simon, Adam, and all those who work on Zotero,

    I appreciate that Zotero is free and, especially considering that, it is well supported. However, documentation has always been a shortcoming and with an ongoing stream of feature changes there should be a proactive stance in making users aware of new features. (I'm not talking about bug fixes.)

    Unfortunately, FireFox seems to have a security update every month. So most of us automatically accept updates to FireFox and most of us have it automatically update plugins. I work with Zotero every day and I wasn't aware of most of what's in the change log. Changes big and small are buried in it. I think it would be to your benefit and to the benefit of your users if you would document the program better and inform users of significant changes to Zotero. One way of doing so would be by periodically emailing your users and telling them about the latest features.

    You've built a fine piece of software and we would all be better off if we knew more about it.

    Best regards,
    Ron Wielage
  • Two comments on this:
    1. Your situation/position in this is more unusual than you seem to think it is. We do get questions/concerns from users when features like this change, but I can not remember any other case where users where so exasperated about a change that is easily reverted/addressed by changing a preference. The usual response to this is "oh, cool, ok". I understand that your needs may be different from other users' and that's fine, of course, but given those, I really do not see or envision any alternative to closely following the Zotero and citeproc changelogs, which we point you to above. Especially the Zotero changelogs are non-technical in language and if you don't understand a described change you can ask here.

    2. That said, I don't think Zotero does a great - or even decent - job of announcing new features (though journal abbreviations - with a screenshot of this specific preference - were, in fact, covered in the 4.0 announcement on the blog):
    Partly because of that I started on which I do plan to cover anything major that happens (though I certainly won't mention minor changes in citeproc behavior, so you'll still be stuck with checking changelogs). See e.g. my posts on the 4.0 release
    for an example. Obviously that's an unofficial blog, no guarantees etc. but you may find it useful. (While I do wish the official blog were more active, I don't think Zotero should send out hundreds of thousands of e-mails on every update).
Sign In or Register to comment.