LSE site translation not working

URL not working with DOI translator

http://blogs.lse.ac.uk/impactofsocialsciences/2014/05/22/thomas-piketty-data-make-it-open/

DOI translator
Zotero 4.0.24.1
Win 7 Pro
Firefox 34.0 plugin
Report ID: 1098005820
  • That's because there's not actually a DOI on that page. DOIs have very lax restrictions on length and characters, so all Zotero can do is make educated guesses about what is a DOI. In this case, there's a filename 10.48.45.jpg-180x138.jpg that it thinks could be a DOI (because of the 10. in the beginning).
    There really isn't much we can do about false positives like that.
  • adamsmith: That's actually not quite the issue. We require a "/" in DOIs, so those image filenames aren't what's triggering this. If you switch to the comment view down a bit on the right-hand side, there's a comment with a link to a Wiley article that includes a DOI in the URL, but the comment is truncated with an ellipsis in the middle of the URL. (It's plain text at that point. This page only includes a link to the comments view, where that comment, and the URL it contains, is provided in full.)

    But still not really a good fix here, and even if it hadn't been truncated it would still be a DOI for a different article.
  • I have forwarded your comments to the site admin. Thanks for your help.
  • They're not really doing anything wrong, though. The DOI with elipses is a bit odd, but as Dan says, at that point you're already looking at a different article.

    Dealing with sites for which we don't have a dedicated translator is never going to be perfect. We do think--and have specific plans to--improve results for pages where multiple generic translators are detected (the blog also has some embedded metadata which Zotero can read), but that will likely still take some time.

    (@aurimas - if you get to this before me: we should probably add the property="article:author" type metatags, though I'm not sure where that vocab comes from.)
  • edited December 14, 2014
    article:author is part of OpenGraph (og). It gets a little complicated and I'm not sure why they decided to subdivide the vocabulary this way. We can probably work around it by assuming it's all og. I don't see any incompatible overlap between different types of objects.
Sign In or Register to comment.