rft.tpages in COinS not took into account (Zotero 2.1b2)

I am importing some refereneces from COinS embed in HTML pages. The rft.tpages tag (indicating the total number of COinS) is not imported by Zoetro (using 2.1b2).

  • Confirmed, ticket opened, patch submitted (https://www.zotero.org/trac/ticket/1747). This was apparently overlooked. Let me know if you notice that any other parts of the OpenURL ContextObject spec are not implemented.

    This is minor and should certainly make it into Zotero 2.1 final.
  • I am still working on books (I will work on over item type in the next days).

    There is the same problem when importing a book from a RIS file. The total number of pages is defined with tag SP but is not imported by Zotero.

    When importing a bib file, the pages tag is correctly imported by Zotero.

  • edited November 16, 2010
    The total number of pages is defined with tag SP
    The specification does not state that SP can be used for the total number of pages.
    When importing a bib file, the pages tag is correctly imported by Zotero.
    It is debatable whether this is correct, but we haven't gotten complaints.
  • Since no meaning is defined for SP or page range in book records and RIS doesn't define a tag for number of pages, I'm inclined to think that we should interpret SP as number of pages for books.

    @larmarange: What made you think that SP was correct for number of pages? Are there other programs that put it there? It would be good to see if this is common practice.
  • I'm inclined to think that we should interpret SP as number of pages for books.
    If SP!=EP in a book record, I would definitely disagree. If there was no EP or if it was identical to EP, I'd be more open to it. I, like you, am not opposed to going contrary to spec if we were doing something that was "popular". I don't know if that is the case here: I've probably seen more records typed as book, when they should have arguably been typed as chap.
  • edited November 16, 2010
    There is a puzzling line in the RIS translator, RIS.js:560-561 if(item.pages) {
    if(item.itemType == "book") {
    addTag("EP", item.pages);

    This line of course is never executed, since book items don't have an "pages" attribute, but it looks like someone wanted to use EP for number of pages.

    [Edit: This is in the export function of the translator.]
  • When creating export with EndNote, EndNote used the SP tag to specify the total number of pages for books. I didn't have Reference Manager, so I can not test with this software.

    However, EndNote and ReferenceManager are both developed by Thomson Reuters.
    Unfortunetaly, in the example section of RIS specifications, there is no example with a book.
  • When creating export with EndNote, EndNote used the SP tag to specify the total number of pages for books.Confirmed. And they do not have an 'EP' tag. I'd still be conservative & only import this when EP is not present.
  • Please go to http://github.com/ajlyon/zotero-bits/raw/master/RIS.js and save the file to the translators directory of your Zotero data directory (http://www.zotero.org/support/zotero_data).

    Restart Zotero and see if the number of pages imports correctly.

    noksagt, others: What about export? Should we abuse the spec ourselves and export with numPages as SP?
  • @ajlyon:

    I saved RIS.js in my translators directory but number of pages was still not imported.

    About export, maybe it would be a good thing to make a test with Reference Manager. If both Reference Manager and Endnote are using SP tags, it would be a good idea to export numPages as SP.

    The specifications are unclear because they don't explain how to export the number of pages of a book. So, if Thomson's software are using this tag, we can consider that it is correct to use SP tags for books' total number of pages.
  • Can you post an example RIS file with several book items that use SP for numPages? Just export something from Endnote and paste it here, so I can test.
  • So, if Thomson's software are using this tag[..]
    Fortunately, we have decided to go by the spec in most cases. There are hundreds of products that use RIS, and observing the spec is the only way to even hope for compatibility. As it turns out, Endnote has little regard for the spec and violates it in unpredictable ways in each version; it's better to stick to the rules than bend every which way to support non-standard and undocumented behaviors. Specifications matter.
  • I'd also normally argue to follow Postel's Law & just follow the spec, but (i) Reference Manager is essentially the reference implementation of RIS and (ii) it is reasonably unlikely to break things if we include extra information.
  • Some books refrences exported with EndNote :

    TY - BOOK
    AU - Bachelard, Gaston
    CY - Paris (FR)
    ET - réimpression de la 7e édition Quadrige de 2006 (première édition en 1934)
    PB - PUF
    PY - 1934
    SN - 2-13-054249-2
    SP - 183
    ST - Le Nouvel esprit scientifique
    TI - Le Nouvel esprit scientifique
    ID - 501
    ER -

    TY - BOOK
    AU - Bourdieu, Pierre
    AU - Chamboredon, Jean-Claude
    AU - Passeron, Jean-Claude
    CY - Paris-La Haye (FR)
    ET - 2e édition révisée, première édition en 1968
    PB - Mouton
    PY - 1973
    SP - 357
    ST - Le Métier de sociologue : préalables épistémologiques
    TI - Le Métier de sociologue : préalables épistémologiques
    ID - 519
    ER -

    TY - BOOK
    AU - Durkheim, Émile
    CY - Paris (FR)
    ET - 5e édition Quadrige, 1895 pour la première édition
    L4 - 4059185408Durkheim 1990.pdf
    PB - PUF
    PY - 1990
    SP - 149
    ST - Les Règles de la méthode sociologique
    T2 - Quadrige
    TI - Les Règles de la méthode sociologique
    UR - http://gallica.bnf.fr/ark:/12148/bpt6k20762v
    ID - 518
    ER -

    TY - BOOK
    AU - Hogarth, James
    CY - Copenhague (DK)
    L4 - 2600471297Hogarth 1977.pdf
    PB - OMS Bureau régional de l'Europe
    PY - 1977
    SN - 9290202238
    ST - Vocabulaire de la santé publique
    T2 - La Santé publique en Europe
    TI - Vocabulaire de la santé publique
    UR - http://whqlibdoc.who.int/euro/phie/WHO_PHIE_4_fre.pdf
    VL - 4
    ID - 411
    ER -

    TY - BOOK
    AU - Lakatos, Imre
    CY - Paris (FR)
    ET - édition française de 1994 traduite à partir de l'édition anglaise de 1986
    PY - 1986
    SN - 2-13-045599-9
    SP - 268
    ST - Histoire et Méthodologie des sciences
    TI - Histoire et Méthodologie des sciences : programmes de recherche et reconstruction rationnelle
    ID - 506
    ER -

    TY - BOOK
    AU - Mann, Jonatahn M
    AU - Tarantola, Daniel J M
    CY - New York City, New York (US)
    PB - Oxford University Press
    PY - 1996
    SP - 616
    ST - AIDS in the World II
    TI - AIDS in the World II
    ID - 450
    ER -
  • Those records import correctly for me with the number of pages filled, using the RIS translator I posted. I doubt it matters, but I'm using Zotero 2.1b2.
  • edited November 20, 2010
    Some more remarks about Report type:

    When importing from a COinS, Zotero import the number of pages if it is declared in a rft.pages tag but not if it is in a rft.tpages tag.

    There is no specific tag in COinS to specify the type or the number of a report, but COinS accept a rft.series tag fand a rft.volume tag for a report. These two tags are not imported by Zotero when rft.genre=report.

    About importing/exporting Report in RIS format : Zotero do not export nor import the type of report.

    EndNote uses the M3 tag to specify the type of Report. Specifications are unclear about the M3 tag :
    Miscellaneous 3. This is an alphanumeric field and there is no practical limit to the length of this field.
    It's still the same problem that the RIS specification is not precise about specific fields.

    Best regards
  • About M3 tags:

    Currently, Zotero use M3 tag for exporting/importing DOI of a journal article .

    This tag is used by EndNote to specify the type of article, a specif tag (DO) existing for DOI. In RefMan Export style for ReferenceManager (ftp://support.isiresearchsoft.com/RefMan/Styles/Refman Export.os) the DO tag is also present.

    The RIS specification don't explain how to specify a DOI.
  • Regarding RIS:
    Based on my review of the existing translators, it looks like EBSCOhost is the only one for which we currently look for DOI in the DO tag. If Endnote and RefMan use DO, I suppose we can look for it there as well.

    As for concurring with Endnote and importing report type from M3 and exporting it there too, and importing and exporting DOI only in DO, I'm not sure. Are there any other good sources of RIS data we can compare with? If we can do it without breaking compatibility with other data sources that interpret the vague standard differently, I'm fine with it.

    Regarding COinS:
    My fix for rft.tpages has been submitted, but it hasn't landed yet, since it affects the core Zotero code (OpenURL -- and therefore COinS -- is in the core code).

    What Zotero fields should rft.series and rft.volume map to in Zotero? Right now, they are being discarded since report items don't have corresponding fields. Fixing this is easy, but I'm not sure what behavior we want to see.
  • The tpages issue has been fixed in the Zotero trunk and should work in 2.1b3 when it is released (https://www.zotero.org/trac/changeset/7347).
  • edited November 21, 2010
    BTW, interesting that Endnote exports an ET tag in RIS for edition info. ET is not part of RIS specs, and Zotero apparently imports/exports nothing for the edition field. Could anything be done about this?
  • There's currently a discussion on the zotero-dev list on changes to RIS.js to improve Endnote support. I'm not sure if the ET issue in particular has been raised or addressed, but hopefully the current discussion/proposal will lead to improved support for Endnote's idiosyncrasies. See http://groups.google.com/group/zotero-dev/t/ad55d89ed54a0a91
  • edited November 21, 2010
    Thanks! (and I see ET is included)
Sign In or Register to comment.