• Buggotea, some philosophical thoughts.

      Wednesday, 01 Aug 2007 - 09:51 GMT

      As many of you know, one of the weaknesses with connotea is it’s inability to recognise two bookmarks that have the same doi but different url’s as being the same object. Most academic articles can be found at different locations on the internet, the journal page, an archive, the authors page, but connotea at the moment does not try to unify these different locations into one abstract article item within connotea.

      If you bookmarks an author’s version of the article and you come back to connotea are you more interested in the article itself or do you also expect to be returned to the specific location that you bookmarked that article at?

      This is related to how we go about engineering the connotea database to to resolve this issue, at a first pass.

      I’d love to hear your thoughts.

      Last updated: Wednesday, 01 Aug 2007 - 09:51 GMT

      • Comments

        • Date:
          Wednesday, 01 Aug 2007 - 10:21 GMT
          Pedro Beltrao said:

          I prefer to have a primary link resolve to the publishers website.

        • Date:
          Thursday, 02 Aug 2007 - 04:28 GMT
          Bill Hooker said:

          I got into the habit of only marking the PubMed page back when Connotea was still having trouble understanding publishers’ pages, let alone author’s own versions. I’ve since noticed that Connotea finds the publisher’s DOI for me, which I really like—and I find both (PMID and publisher’s DOI) links useful. If Connotea can figure out that different links point to the same article, perhaps it would be best to have all of those links displayed no matter which one the user bookmarked from?

        • Date:
          Thursday, 09 Aug 2007 - 08:23 GMT
          Ross Gayler said:

          I would prefer to have Connotea display the DOI and all other URLs (perhaps ordered by recency of addition). I do not have an academic affiliation and DOIs almost invariably point to subscription-only sites that I can’t access or would have to pay more than I could justify for a single paper. I need the DOI to cite for the benefit of others with affiliations and the other URLs so I can actually obtain the paper.

        • Date:
          Thursday, 09 Aug 2007 - 17:05 GMT
          Jennifer Forman Orth said:

          Apologies if this is too off-topic (or has been addressed elsewhere), but what about the opposite case: when two references share the same URL? If I am trying to build a reference database, and I have some references I wish to add where the only URL that refers to them is a single page listing all of an author’s publications, when I add the second reference, it just writes right over the first one without even telling me!

          I realize this might be an unintended use for Connotea, but it seems like for a reference db to be useful it should not be restricted to a subset of possible entries…i.e. I want to share my references online with a group, and I don’t want to have one place to look up “newer articles with DOIs” and a second place to look up “older articles and obscure ones without DOIs or unique URLs.” The “U” in URL and URI != “Unique” :-).

        • Date:
          Wednesday, 15 Aug 2007 - 21:25 GMT
          Ross Gayler said:

          I agree with Jennifer. Connotea is most useful if it is a personal one-stop-shop for references, so it has to be able to cope with items lacking a DOI or well-behaved URL.

        • Date:
          Thursday, 16 Aug 2007 - 10:40 GMT
          Ian Mulvany said:

          To refer to Jeffier’s point, there is another case where being able to deal with non-url recourses is important, and that is when importing a bibliography that you might want to share.

          We have been discussing this as it is something we really have to support. I think we are going to fix buggotea first and then deal with the reverse case after that.

        • Date:
          Saturday, 25 Aug 2007 - 04:06 GMT
          Mohsen Sadatsafavi said:

          You can have the best of both worlds! The URL should locate to where it indicates. If someone has bookmarked a Pubmed abstract s/he should be always redirected to that page. However, if one can query links on their doi or pmid, one should be able to retrieve all bookmarks pointing to that article, regardless of their URL. This can happen using a new filter for doi. For instance, the following URL (possible doi in MD5 code)
          http://www.connotea.org/doi/10.1186/1471-2458-6-158

          should retrieve all bookmarks for this article, from Pubmed to the original publisher website to pubmed central etc.

        • Date:
          Monday, 10 Sep 2007 - 16:59 GMT
          Andrew Schamess said:

          I agree with Moshen that both are useful. For my own purposes, it’s helpful to be able to retreive the article on my institution’s proxy server, so I want to save the url where the article is located for me.

          But that defeats the “social networking” aspect of Connotea. By having a common identifier such as PubMed ID for all articles, you can find out who else has bookmarked that article – and explore any similar bookmarks s/he might have made public, form communities of interest, etc.

          The social networking aspect of Connotea could make it a much more useful resource. Imagine if you could search for the most popular (most bookmarked) articles on a topic. Say that I wanted to find out more about neuropsychology. I could go to Connotea and read the ten most popular articles. One would assume that any article bookmarked by, say, 100 people, could be considered an important and widely read work in a given field.

          I understand that work is underway on a Firefox extension that would allow Zotero references to be uploaded to Connotea. Since Zotero automatically collects standard reference data on all articles saved via the “import reference” function, this would provide a great opportunity to link articles by common fields other than the url. Anything uploaded from Zotero would contain the PubMed ID. Connotea would then need to be structured to link articles with the same PubMed ID - or whatever field was considered an appropriate common denominator.


Search blogs

web feed Want a blog?

Submit this post to

Advertisement