Sonny Software

Being able to access my CiteULike library from within the Bookends Browser, and selectively import references along with the attached PDFs, is for me one of the hottest features of Bookends. Because I often can't remember what I've imported in the past, I often end up with duplicates. Bookends' Remove Duplicates… feature is pretty powerful, so remedying the problem isn't exactly rocket surgery, but if Bookends could check the Citeulike-Article-Id and grey out items that are already in the frontmost database, that would in my mind be a great substitute for full synching ability. Would such a feature be feasible?

Hi,

Bookends already does something like this (marks the ones you have in yellow) for PubMed and JSTOR. That's because they have unique identifiers that Bookends can look for. I don't believe CiteULike records have anything similar. But some CiteULike records do have dois. I'll think about the possibility of using this.

Jon
Sonny Software

I like this idea. It will be implemented in Bookends browser for JSTOR and CiteULike (the latter requires that the reference has a doi. If it does not, Bookends will not recognize it as being in the open database).

Jon
Sonny Software

Jon wrote:I like this idea. It will be implemented in Bookends browser for JSTOR and CiteULike (the latter requires that the reference has a doi. If it does not, Bookends will not recognize it as being in the open database).

Wonderful! Thanks, Jon.

Another question about the Bookends Browser and CiteULike: Is there any way to make Bookends pull in page numbers when that information is included in the entry on CiteULike? At the moment, I'm adding it manually after importing references.

I'll look into that.

Jon
Sonny Software

Bookends already does something like this (marks the ones you have in yellow) for PubMed and JSTOR. That's because they have unique identifiers that Bookends can look for. I don't believe CiteULike records have anything similar.

Just a clarification, Jon: I think every CiteULike item has a unique ID. A typical BibTeX output has something like this in it:
citeulike-article-id = {5173786}

while RIS output for the same article has:
ID - citeulike:5173786
L3 - citeulike-article-id:5173786

Hi,

The pages import from CiteULike will be fixed in the next update.

The CiteULike unique identifier is not included in the COinS embedded object, which is what Bookends browser is importing. So that can't be used to identify duplicates in the database. The only reliable tag that I can see is the doi, and Bookends will look for that (unfortunately, the doi is included in most scientific articles but not for many in the humanities or for books, so it only works in some cases). If you would like for CiteULike to include the unique identifier in their COinS objects, please write to them and make this suggestion. I'm sure they'd like feedback on how their work with COinS is being used.

Jon
Sonny Software

Jon wrote: The CiteULike unique identifier is not included in the COinS embedded object, which is what Bookends browser is importing… If you would like for CiteULike to include the unique identifier in their COinS objects, please write to them and make this suggestion. I'm sure they'd like feedback on how their work with COinS is being used.

Thanks for clarifying that, Jon. I'll do as you suggest. It's my experience that the CiteULike people are very responsive, just like someone in these parts. Thanks also for taking a look at the page numbers issue.

Hi again,

The CiteULike people have already responded favorably. But they want to know exactly what format would suit Bookends' purposes best. Here's the message:

Easy to do, but how? With a DOI there's
rft_id=info:doi/10.1232/121212121
I could do
rft_id=info:citeulike/1233221
if there's no DOI, or use a completely different (custom field), e.g.
cul_id=23131313
I suggest you need to ask Bookends what will suit.

What should I say?

Thanks again for being so responsive on this.

Here's the thread in case you want to get involved directly:
http://www.citeulike.org/groupforum/1467

More from the CiteULike people:

I've had a look at the specs , and I think the proper way to do it is:
rfr_id=info:sid/citeulike.org/citeulike (already there)
rft_dat=article_id/12242424
Posted by thegoose on 2009-07-19 10:20:23. Reply.
...but another interpretation says we can have multiple rft_id
rft_id = info:doi/10.1126/science.275.5304.1320
rft_id = info:pmid/9036860
rft_id = http://www.citeulike.org/user/fred/article/123456 ??????

For anyone following this discussion (I hope I'm not the only person who uses both CiteULike and Bookends!), Jon and the CiteULike people are working together directly on this.

The developers of CiteULike and I have settled on a way of providing a unique identifier in the COinS output, so Bookends will identify citations you already have in your database for you in the next update. This will apply only to CiteULike citations downloaded into the next version of Bookends (and later), because that unique identifier will have to be imported as well (and will be stored in the Notes field). So you'll have to wait for 10.5.6 to see this work.

Jon
Sonny Software

Thanks a bunch, Jon. That's a big, big help.

Looking forward, any chance in a future version of Bookends storing CiteULike account information? If you haven't used CiteULike, you may not know this, but when you're logged in and view your own library, PDF icons appear for any articles for which you own a PDF, and you can download them (but nobody else can see them). If Bookends knew who I was in CiteULike it could download the PDFs when available along with the references, rather than trying to do so from Google Scholar and the usual places (where the PDFs usually aren't available).

I'm painfully aware that CiteULike isn't something that many Bookends users are excited about (I like it because, using the CiteULike bookmarklet, I can store references from the computers in university libraries when I visit them and don't have, or aren't allowed to hook up, my own computer), so I'm more than satisfied for now with the coming enhancements in 10.5.6. I just wanted to put it in the back of your mind that there's more that could be done should the time and opportunity present themselves to make the Bookends-CiteULike integration more seamless.

You can already download the pdfs you have in CiteULike (if you can actually open them when in CiteULike, which I assume you can). Just navigate to the pdf, then drag and drop the pdf proxy icon (the dotted square) onto the Bookends database window. The pdf you are viewing will be downloaded and attached to the selected reference.

Jon
Sonny Software