MS integration Bibliography... "retrieved" unwanted

tibr
Hi,
Zotero is really a great tool, but i'm having a problem with MS word integration: in the bibliography, APA formatting, i get this kind of info at the end of each reference:

Retrieved April 3, 2007, from http://dx.doi.org/10.1007/BF00340571.

I'd like not to have this info, what can i do ? is this a bug or have i missed something ?
does anyone have the same problem ?

I'm using it with a macbook, with MS word 2004 for mac.

thanx a lot for your help.

tim
«1
  • dstillman
    "Retrieved...from" is added when a reference has entries in the URL and Accessed fields. With the default styles, you'd have to remove those values to get them not to show up in generated bibliographies.

    Now, one question open for debate is which translators should put data into those fields (as opposed to a web link attachment). It's possible that the URL field should be reserved for canonical URLs from the original provider—for example, the URL of a New York Times article—rather than the URL of an article on one of many sites that offer the full text content...
  • bdarcus
    This gets into some tricky issues about identity and location. I personally think this field should be for in essence perma-link URLs: URIs that can be used as rdf:about values for web-accessible resources.

    From that standpoint, any web document applies. This includes NYT articles and such.

    But journal articles and books are much more tricky. I think this really bears more in-depth discussion as a part of the larger questions of URIs and identification.

    Will think on this some more ...
  • ransomca
    Any progress on this front? I'm writing a dissertation and I have hundreds of references I have to go back and delete this stuff from. And I better not update the bibliography because if I do, I have to go back through and do it all over again. Why can't we choose which fields get outputted to the reference list? Or only have certain reference types (ie, web pages) list the url information? (Mac OS X Tiger, Word 2004, Zotero Word plug-in)
  • CB

    "Retrieved...from" is added when a reference has entries in the URL and Accessed fields. With the default styles, you'd have to remove those values to get them not to show up in generated bibliographies.
    APA style mandates retrieval information only for electronic sources. So to get non-electronic sources to appear in correct APA style with Zotero, it's currently necessary to delete data (which may be required for other purposes). This strikes me as odd.
  • bdarcus
    APA style mandates retrieval information only for electronic sources. So to get non-electronic sources to appear in correct APA style with Zotero, it's currently necessary to delete data (which may be required for other purposes). This strikes me as odd.
    In a few years, it will seem "odd" to talk about "electronic sources" at all.

    The conceptual issue is that a URI can be used in this context either to identify the source per se, or a copy of the source. The way the processing works now is that Zotero is assuming the content of the URL field refers to the former, while many people use it to store information about the latter.

    Somehow we need a way to make this distinction crystal clear in the UI. I think the basic distinction between a URI as identifier for the source and as a link to a copy is fine, and supportable logically. It's just not totally intuitive for users (well, nor is the distinction between URL -- an essentially meaningless concept -- and a URI).
  • CB
    I'm pretty ignorant about the issues you're talking about here, and accept the point about the limited longevity of distinguishing between non/electronic sources. But isn't the reference-display issue raised by the OP (and reiterated by me) separable from that of uri semantics? (Not to denigrate the importance of the conceptual issues, as I know how decisions on things like this can come back to bite later).

    If someone wants to create an APA-style bibliography, today, using Zotero, they either have to delete a lot of potententially-useful data from their Zotero db, or have to do lots of tedious manual editing of Zotero's output.
  • bdarcus
    If someone wants to create an APA-style bibliography, today, using Zotero, they either have to delete a lot of potententially-useful data from their Zotero db, or have to do lots of tedious manual editing of Zotero's output.
    No, they don't have to delete it. They do have to move the content (to a link).

    That itself might be tedious, but it's at least doable.
  • dstillman
    No, they don't have to delete it. They do have to move the content (to a link).
    And that process could perhaps be automated by a right-click option to "Convert to Web Link Attachment".

    It's also possible that there are translators that are still putting the URL in the URL field when they should be creating web links.
  • CB

    No, they don't have to delete it. They do have to move the content (to a link).

    That itself might be tedious, but it's at least doable.
    But I'm still not clear on why it's necessary at all. APA formats for, say, standard books and journal articles, just don't include any URL-type information. So why should either a "canonical" uri (don't know if this is the right terminology) or any other link be there? I can see why there needs to be a decision on which url field to include for electronic references, but for the (currently) most common types of reference there's no issue.
  • dstillman
    APA formats for, say, standard books and journal articles, just don't include any URL-type information.
    Would that still be the case if the book or journal article was distributed solely online and the URL contained its location? Perhaps so, in which case it may be appropriate for the style to determine the Retrieved From line based on the item type, but it seems that the canonical URL would be an essential part of a citation for such a source.

    However, if there are translators that are putting database URLs into the URL field for books or journal articles, it's quite likely that those are just bugs that need to be fixed, as any URL that points to just a listing or copy of a source should be a web link attachment. If those translators were fixed, even if manual conversion were necessary to conform to a style, it would probably be a rare occurrence.
  • CB

    Would that still be the case if the book or journal article was distributed solely online and the URL contained its location? Perhaps so, in which case it may be appropriate for the style to determine the Retrieved From line based on the item type, but it seems that the canonical URL would be an essential part of a citation for such a source.
    Yes, you're quite right. I was thinking of ebooks as being a different type from a print book, but actually the latest APA (5th ed.), dating from 2001, doesn't explicitly include such a type. It does talk about a variety of 'electronic sources', but obviously they are a set of moving targets, and the publications manual is already looking a bit dated. I can see now that Zotero's 'book' (for example) will have to do for both e- and print-books, and therefore the 'Retrieved from ..' section will have to depend on the existence or otherwise of a canonical URL.

    However, if there are translators that are putting database URLs into the URL field for books or journal articles, it's quite likely that those are just bugs that need to be fixed, as any URL that points to just a listing or copy of a source should be a web link attachment. If those translators were fixed, even if manual conversion were necessary to conform to a style, it would probably be a rare occurrence.
    Yes, I see where you're coming from now. Most of the translators do seem to put in a URL as well as a link, especially those culling journal articles from scholarly databases. Should these be reported somewhere?
  • Matthias
    Personally, I find it also confusing that I'm not allowed to put regular URLs (i.e. "non-identifying" URLs that point to copies of the source) into the URL field. I think that it would be best to rename the current URL field to something more specific, which would help the user to understand that the field is meant to hold an unique identifier (URI) and not just a regular URL.
  • bdarcus
    I agree Matthias. I think the solution is probably to introduce some explicit notion of copy (in frbr terms, item) into the UI. So there are local copies on the file system, and there are global copies identified with URIs.

    In essence, a link is a relation between an abstract source (document) and a copy of it. That needs to be make clear.
  • dstillman
    Most of the translators do seem to put in a URL as well as a link, especially those culling journal articles from scholarly databases. Should these be reported somewhere?
    Yes, please. You can post them here, and I'll make sure they end up in a Trac ticket.
  • CB
    One situation that arises a lot now, probably because of the transitional phase we're in, is journal articles that have their primary identity as paper articles, but also have a canonical net reference (eg. doi).

    So isn't the situation conceptually that we have three types of web references: canonical source (ie. for 'true' electronic publishing), canonical location (but with the true source being elsewhere, eg. on paper), and copy?

    Presumably we do want to distinguish between the former two (because canonical sources need to be in a bibliography, but in general, at present, canonical locations don't; and also between the latter two -- because there really is a difference between a casual URL copy and something like a doi (even for material primarily published on paper).
  • bdarcus
    CB: I think the way to think of this is to go back to FRBR. A journal article has a DOI. That DOI is an identifier for a particular expression of a work.

    In turn, that expression has three different manifestations; aka formats: print, PDF, HTML. Each of those in turn may have their own identifiers, and they may include URIs.

    It so happens that with many "web resources" there is only one manifestation, which happens to be served up when you enter the URI for it in your browser.
  • CB
    bdardus: This isn't a field I've looked into or thought about that much (I had to look up FRBR), so I'm aware that anything I say here might be naive in the face of prior art.

    The logic of the scheme you've set out is clear enough, but it looks a little like a vision of the future rather than an exact description of the current situation. On the face of it, as things stand today, isn't there a difference between a large publisher's print version of a journal and, say, a casual printout of an ejournal article?

    Perhaps it doesn't really matter. The only reason it seems to at all is the practical issue of automatically generating bibliographies, which, in most current styles in use, really do distinguish between item types where uris are and aren't reported. I expect uri's will be mandated for all contemporary sources in most styles before too long, but "before too long" might be 5, 10 or 20 years, and there will be a large number of bibliographies generated in that time!
  • bdarcus
    The logic of the scheme you've set out is clear enough, but it looks a little like a vision of the future rather than an exact description of the current situation. On the face of it, as things stand today, isn't there a difference between a large publisher's print version of a journal and, say, a casual printout of an ejournal article?
    Sure.
    Perhaps it doesn't really matter. The only reason it seems to at all is the practical issue of automatically generating bibliographies, which, in most current styles in use, really do distinguish between item types where uris are and aren't reported.
    In my experience, the fact of whether a source has a URI makes a trivial difference in formatting, as it should.

    But yes, bibliographic styling is a pretty arcane process that ought to get rationalized as scholarly finally catches up with the web. It might indeed take a few decades though ;-)
  • CB

    post them here, and I'll make sure they end up in a Trac ticket
    Dan, the translators for the following sites wrongly fill the URL field for print journal articles:

    Blackwell Synergy eg. http://www.blackwell-synergy.com/doi/abs/10.1111/1467-9507.00059

    Ingenta Connect, eg.http://www.ingentaconnect.com/content/jbp/is/2005/00000006/00000003/art00009

    Annual Reviews (http://arjournals.annualreviews.org.)
    This one also inserts the article abstract in "Extra" instead of "Abstract"

    I've no idea how to point to a translator beyond giving you the site/URL.

    Also, Zotero's interaction with my University library OPAC (http://library.uq.edu.au/screens/opacmenu.html) is patchy: with books, sometimes the Zotero 'book' icon shows, sometimes not. There's no obvious pattern to it. Is there a way of finding out what means Zotero is using to access the book info when it does find some?
  • scot
    I've wondered at this last point, too. Is there (could there be?) some way for the user who wants to make a proper bug report or enhancement suggestion to know what translator he is reporting on? ("ALEPH" translation is provided by the MARC translator, etc.)

    The EBSCOhost translator puts URIs for the database record into the URL, even when it's only a database record and no full text articles are offered or linked to.
  • giclin
    I'm wondering what's the situation of this problem now.

    I need to write in APA format and it's been very annoying to have the doi information in the bibliography. I used the database mentioned above (Ingenta Connect, EBSCOhost...) to capture references and PDFs a lot. So I have many references showing "Retrieved...".

    I agree with all the issues discussed above, but to me the most important point is that while most of my references are journal articles (not online only ones) and I don't need to show the doi info., and I believe it is the case for most people writing academic (APA) papers/dissertations, so shouldn't this somehow be the default option, and Zotero can keep working on how to give options for people to decide themselves or how to separate the different URL fields?
  • scot
    There is a preference for this on the Export tab in Zotero's pref dialog. It works correctly for me in Chicago. Seems to be smart, too. It cites the URL if there is no page range given, and omits it if there is. Is that what you get?
  • giclin
    yes, i don't have problems with the URL. i set up it as you said.

    But i still get the DOI showing in many of the references. is there a way to turn it off?

    thanks!
  • Tjowens
    APA now includes DOIs for articles in citations.
    ( http://forums.zotero.org/discussion/130/?Focus=8749#Comment_8749 ) Export does not currently allow to DOIs to toggle on and off like URL's and access info. But if there is enough interest/need for this kind of feature we can create a ticket.
  • giclin
    Thank you for pointing me to the link. But looks like it mainly refers to the electric version.
    What we usually encounter though are articles with both electric and print version. If it's not required to report DOI in this case, than I think it might be better to set it as default (not showing DOI), but also allow the option to toggle on or off for each reference, or for a global setting.

    Offering another option for "Electric Journal article" is also a good idea (proposed from that link).

    Just my two cents.
    Thanks!
  • bdarcus
    For sake of argument:

    What could possibly be wrong about including a DOI? It's information that is valuable to readers, and it doesn't take up much space.
  • tonyc2001
    edited April 16, 2008
    Unfortunately, I suspect those that mark my thesis may not think that it's valuable. To them, it's either by the APA manual or it's wrong. As far as I know, 5th ed APA doesn't include DOI unless the journal is electronic only. I'd really appreciate a way to turn the DOI off on the reference list, and maybe turn it on only for a new 'Electronic Journal Article' document type.

    On the subject of tweaking APA style reference lists:
    - there shouldn't be a blank line between items (mentioned in http://forums.zotero.org/discussion/2252/word-plugin-inserts-superfluous-linebreaks-in-bibliography/#Item_13)
    - The volume number needs to be italicised (I see a change for that is underway, I think: https://www.zotero.org/trac/changeset/2560)
    - the reference list needs to be double spaced (though I'd prefer that no style info was passed, leave it up to the document style, also mentioned in http://forums.zotero.org/discussion/2252/word-plugin-inserts-superfluous-linebreaks-in-bibliography/#Item_13)

    Cheers
    Tony
  • Bryan Pfaffenberger
    My understanding of the APA style (based on the APA recommendations of 2007) is as follows:
    • For electronic journals lacking an identical print counterpart AND lacking a DOI, "Retrieved from... + URL" are required.

    • For electronic journals that do supply a DOI but lack an identical print counterpart, the DOI is required and the URL should be omitted. (DOI is a registration service that is intended to solve the problem of changing URLs.)

    • For electronic journals that DO have an identical print counterpart, "Retrieved from..." + URL are NOT required but the DOI is still recommended.
    These rules reflect the realities of current practice, which are, of course, undergoing rapid change. DOIs enable database vendors to avoid the hassle and expense of maintaining a stable URL, but the DOI registration service is a commercial enterprise and smaller publishers may not be able to afford a subscription. And as libraries continue to eliminate print subscriptions while retaining electronic ones, an author's failure to include a DOI for a print-based article may impose additional retrieval costs on readers whose libraries have eliminated their print subscriptions. Novices often have trouble determining which database to search. Providing the DOI eliminates this problem.

    If I've interpreted the current APA rules correctly, they suggest the following:

    • There should be a new source type ("Electronic Journal Article") for online journals that do NOT have a print counterpart. This source type should print the DOI, if available; otherwise, it should print the URL (but I agree that users should be able to suppress this if they wish).

    • The existing "Journal Article" source should print the DOI at the user's option.

  • cmduke
    A related issue I've encountered . . . Even for an electronic copy of a print resource, I can not determine what causes the "Retrieved from URL" to toggle on and off in the bibliography when auto-generated using the MS Word plugin.

    -cmd
  • adamsmith
    cmduke - that depends on the style - what are you using? APA?
Sign In or Register to comment.