IOP (Institute of Physics) journal: DOI, not IOP used, creates error

schmid
When trying to import the following article:
http://iopscience.iop.org/0957-0233/1/9/021/
doi: 10.1088/0957-0233/1/9/021
after several seconds, I get 'An error occurred while saving this item'.

Strange enough, the 'Save to Zotero' symbol says 'DOI', not 'IOP'!
The IOP translator should be used; it seems that detailed data are missing in the crossref/doi database (it is a rather old article).

Zotero 2.0.2 on Firefox 3.6.3, Mac OS X 10.4.11
  • daniel.habersat
    I have also encountered this problem, at different articles on the same website.

    URL: http://iopscience.iop.org/0268-1242/8/6S/004
    Report ID: 2028376271
    Zotero 2.0.2, Firefox 3.6.3, Windows Vista SP2
  • diffenderfer
    I think I have the same problem while importing:
    http://iopscience.iop.org/0031-9155/27/7/001/

    Using zotero 2.0.2 on Firefox 3.5.9, Ubuntu 9.10
  • levbishop
    Any progress on this? I seem to be experiencing the same problem:
    Report ID: 1719647593
    URL: http://iopscience.iop.org/0305-4470/13/2/034/
    Zotero 2.0.3, Firefox 3.6.8, Windows XP SP3
  • ajlyon
    I worked on it again, but I'm still running into some issues with my approach to falling back to RIS data. So no solution yet, but maybe soon.
  • ajlyon
    If you'd like to try my work-in-progress translator, go to http://github.com/ajlyon/zotero-bits/raw/master/Institute of Physics.js and save the file to the translators directory of your Zotero data directory (http://www.zotero.org/support/zotero_data).

    Reload the page and see if it works. It might not yet work.
  • bbolker
    Nope. But it does at least report "Institute of Physics" rather than "DOI" on hovering.

    Thanks for your efforts.
  • blackjancsi
    @ajlyon: hi there, I have a similar problem. I needed a browser restart (not just a page reload) to have the tooltip show "institute of physics" as for bbolker. However, once I click the Zotero icon at the right of the browser address bar, the notification box (with a red frame) comes up, remains empty for about 5-6s, and then disappears (so basically, no error message, and no record added.

    This is the record I am trying:
    http://iopscience.iop.org/0741-3335/37/1/002

    (also, all the others from this volume of the journals are broken).

    Thx for the effort, even if you won't take it up again!

    Janos
  • ajlyon
    As noksagt explained in the other thread, this is in part because we can't get reliable data for IOP. Between RIS, BibTeX and DOI (CrossRef), we could patch something together, but it is relatively hard to make a translator that uses multiple data sources in such a way.

    If they provided quality data in any one format, it would be rather simple to make a translator. Feel free to contact IOP and encourage them to improve their metadata export.
  • noksagt
    Do you have examples of poor RIS data? Removing the DOI part from your translator & relying on RIS made reasonably acceptable entries as far as I can tell...
  • ajlyon
    I remember that being the reason I tried to implement a fallback system. Unfortunately, I don't have test URLs to justify it. I've modified the version on Github as you describe-- let me know if it works for you (noksagt and others).

    But note that the example URL posted by Janos has low-quality RIS-- it has: A1 - A Herrmann et al At the same time, CrossRef gives us the author names in their entirety and the complete page range; RIS gives only the start page.
  • schmid
    Currently no success with IOP at all, e.g. http://iopscience.iop.org/0957-4484/6/4/001/ uses the DOI translator, but gives me Error:
    DOI Translator: could not find DOI Source File: .../zotero/translators/DOI.js Line: 0
    When entering the DOI manually at dx.doi.org, it links to the correct paper, so it should not be an unknown doi.
    Firefox 3.6.12, Zotero 2.0.9, also tried "reset translators&styles'.
  • noksagt
    edited November 30, 2010
    My version of the IOP translator works for me (using RIS data).
    When entering the DOI manually at dx.doi.org, it links to the correct paper, so it should not be an unknown doi.
    The error message could be improved, but the crossref data lacks an article title, which is required for the zotero translator to work.
  • schmid
    Sorry, still no success - is there any trick different than cklicking on the "Save to Zotero" icon in the Address bar?
    My 'Institute of Physics.js' translator has "translatorID":"9346ddef-126b-47ec-afef-8809ed1972ab", "lastUpdated":"2008-04-28 17:50:00". File modification date is Oct. 12, 2010, however, clearly much newer than my Zotero installation.
    So the translator file contents look outdated, but updating from the 'General' tab of the Zotero Preferences does not give me anything newer.
    Any hints are welcome - thank you.
  • ajlyon
    You can try the translator modifications I'm proposing, as I note above:

    If you'd like to try my work-in-progress translator, go to http://github.com/ajlyon/zotero-bits/raw/master/Institute of Physics.js and save the file to the translators directory of your Zotero data directory (http://www.zotero.org/support/zotero_data).

    It works, but the RIS data isn't always that great (in my opinion/experience). It would be good to contact IOP to explain to them that their RIS data is rather low quality (as I described above). Also, the data they've submitted to CrossRef is sometimes incomplete. This puts us in something of a bind-- but we can push the RIS version of the translator (at Github now) to clients if it will bring significant improvements vs. DOIs.
  • schmid
    Hi Ajlyon,
    your work-in-progress translator works much better than the current one, but I agree that the RIS data privided by IOP are far from perfect (e.g. multiple authors in a single author line, no end page given).
    Now I also discovered why the standard IOP translator in zotero does not work - it does not recognize page names like 'http://iopscience.iop.org/0957-4484/6/4/001/': 
    "target":"^http://www.iop.org/EJ/(toc|abstract|search|article)"
    Maybe someone could update this?
  • ajlyon
    The target expression is only part of the problem -- I can submit the translator that I posted here to be pushed to clients, but I'm not sure that it's better than letting clients fall back to DOI, since CrossRef has better data. Do any IOP DOIs work for you?

    My suggestion that you contact IOP was not an idle one-- there are real bugs in their export output, and they may well fix it if we ask them to. I suggest you personally explain the situation to them and/or you ask your librarians to contact their account representatives to explain this-- it's a bug on their end, and it's preventing paying customers from using the database effectively.

    Feel free to point them to this thread and to the Zotero developer list (http://groups.google.com/group/zotero-dev/) if they have questions about what is wrong with their export data.
  • schmid
    Hi Ajlyon,
    I agree, we have to push IOP to provide better data - it can't be their intention to make it difficult to cite their articles and thereby lower their impact factors...
    For the time being, until IOP fixes the problems, maybe a workaround could be something like the following? (in a sort of symbolic Java-like code; unfortunately I am not good at all in JavaScript) 
      try { use_doi_translator }
      catch ( doi_not_working_error ) { use_RIS_data_from_IOP }
  • ajlyon
    I tried to implement an IOP translator with a try-catch fallback from DOI to RIS. It didn't work, and I'm not quite sure why. There's only one translator that currently works that way (kudos to Frank/fbennett), and I hope to replicate it for some other sources that need it, but I haven't had a chance to do so yet.
  • michael_roberts
    Hi,

    I'd like to introduce myself - I work for IOP Publishing and I have responsibility for maintaining our metadata in CrossRef. I'd like to thank schmid for bringing this and other related threads to my attention.

    We have identified the lack of titles (and other metadata) for older articles in CrossRef as an issue and it is something we intend to resolve. We are also aware of the need to improve the RIS export from IOPscience. I work alongside the IOPscience team and I will be asking them to address this situation as well.

    I'd like to mention that there are other sources of metadata to be found as well - IOPscience includes metadata in each article HTML page - information available includes the bibliographic data (journal, volume, issue, start page), title, DOI, and the list of authors (as both a single complete list and individual names).

    I'm not familiar with how Zotero extracts information from an HTML page, but would this data help with the issues you've been experiencing?
  • ajlyon
    Michael,

    Thanks for contributing here. It would be rather simple to rewrite the translator to use the metadata embedded in the page code. See https://www.zotero.org/svn/extension/trunk/translators/Institute of Physics.js to see how the translator currently works-- it essentially crafts a URL to get the RIS for a given article and hands the RIS off to Zotero's RIS translator.

    The meta tags in the page source are enough to not need to consult CrossRef or RIS, so it looks like the IOP translator will have to be rewritten to use them; in fact, we can eliminate one server request by making the switch.

    I won't be able to rewrite the IOP translator for at least a few more weeks (I'm behind on translator reviews and patches already!), but this one should be really easy for anyone with even a little bit of JavaScript knowledge.
  • jmgilligan
    A comment about IOP RIS generation. When I export a citation from an IOP journal (environmetnal research letters) to disk in RIS format and try to import the resulting file into Zotero (2.0.9, Firefox 3.6.13, Windows XP) from the Zotero "import" menu, Zotero pops up a dialog about saving an RIS file, asking which application to use, as though it's downloading the file rather than importing it. After I dismiss the dialog, Zotero hangs and is unresponsive until I close Firefox and restart it.

    If, on the other hand, I export the IOP citation to disk as "Endnote Text" format, and then import that file into Zotero, it imports fine.
  • ajlyon
    I haven't tried to rewrite this to use the embedded metadata, which would be more efficient-- but I did get a working fallback going, using RIS when CrossRef fails.

    Please go to http://github.com/ajlyon/translators/raw/master/Institute of Physics.js and save the file to the translators directory of your Zotero data directory (http://www.zotero.org/support/zotero_data).

    It should start working again. If this works for you, please post here so that I can submit this change to be pushed to all users.
  • michael_roberts
    Following up on my previous post, I can confirm that we have completed the re-deposit of metadata with CrossRef for the articles that were missing information like title, etc.

    As an example, the XML returned by this request now includes the title:
    http://www.crossref.org/openurl/?pid=zter:zter321&url_ver=Z39.88-2004&req_dat=usr:pwd&rft_id=info:doi/10.1088/0957-4484/6/4/001&format=unixref&redirect=false

    I am hopeful that you should now see much better results when importing IOP Publishing articles into Zotero.

    Regards,
    Mike
  • ajlyon
    Glad to hear that the CrossRef data has been fixed. Thanks for your diligence!
Sign In or Register to comment.