Hi,
Just thought of sharing a few things from my limited experience...
While majority of the new articles from pubmed downloaded the full text, the following did not - yes I could do it manually from the OpenURL page through the link-out.
http://www.ncbi.nlm.nih.gov/entrez/quer ... s=10674721 (JStage Link out)
http://www.ncbi.nlm.nih.gov/entrez/quer ... s=16421047 (meta press link out)
http://www.ncbi.nlm.nih.gov/entrez/quer ... s=17135575 (journal of leukocyte biology full text)
http://www.ncbi.nlm.nih.gov/entrez/quer ... s=17124428 (karger full text)
Hope it helps to fine tune the software in the future updates.
Failed full text downloads
Hi,
Actually, it's already pretty fine tuned.
I've gotten Bookends to work with most of the major publishers, and a few journals that don't belong to a larger group (e.g. Science). But I'm not going to try to have Bookends work with individual journals that have their own idiosyncratic servers or publishers of journals that generate little interest. To do so would be quite wasteful. As you note, the majority of articles are directly accessible (in my own random tests, I found that about 80% of journal articles I had access privileges for were found, but YMMV).
If people have particular publishers that have a reasonably large number of journals thay they want to access directly, or one particular popular journal in an area I have omitted, you can bring them to my attention and I can see if anything can be done.
But as you noted, it's not too difficult to download the 20% or so that fail the old fashioned way. You don't need OpenURL, either. It's faster to use Refs -> PubMed -> Full Text (If Available).
Jon
Sonny Software
Actually, it's already pretty fine tuned.
I've gotten Bookends to work with most of the major publishers, and a few journals that don't belong to a larger group (e.g. Science). But I'm not going to try to have Bookends work with individual journals that have their own idiosyncratic servers or publishers of journals that generate little interest. To do so would be quite wasteful. As you note, the majority of articles are directly accessible (in my own random tests, I found that about 80% of journal articles I had access privileges for were found, but YMMV).
If people have particular publishers that have a reasonably large number of journals thay they want to access directly, or one particular popular journal in an area I have omitted, you can bring them to my attention and I can see if anything can be done.
But as you noted, it's not too difficult to download the 20% or so that fail the old fashioned way. You don't need OpenURL, either. It's faster to use Refs -> PubMed -> Full Text (If Available).
Jon
Sonny Software
Jon,
I agree with you that there are just too many publishers and apparently a lack of standards in how full text articles are presented to the user. In that light. In that light I do commend your job. And I am not unrealistically asking it should work with every eerie journal out there.
If you notice, two of the four example links above gives public free full text access. Doing,
Refs»Pubmed»Full text(if available) with the Journal of Leukocyte Biology article opens the PDF in browser! So in other words BE is able to make safari fetch it correctly, but unable to download it to attachments folder? I found that particularly odd and suggested perhaps it could use little more fine tuning.
The first link (JStage) also gives public access to the PDF, but the article.pdf is one click away after the Refs»Pubmed»Full text(if available) maneuver. Again I wonder if bookends can open the full text page without a problem and the PDF link is right on that page, how come it couldn't download it.
Perhaps you have perfectly good reasons for such behaviour and as a mere end user I may be just oblivious to them – but I thought I would give my 2¢ with the hope it'd be useful.
I agree with you that there are just too many publishers and apparently a lack of standards in how full text articles are presented to the user. In that light. In that light I do commend your job. And I am not unrealistically asking it should work with every eerie journal out there.
If you notice, two of the four example links above gives public free full text access. Doing,
Refs»Pubmed»Full text(if available) with the Journal of Leukocyte Biology article opens the PDF in browser! So in other words BE is able to make safari fetch it correctly, but unable to download it to attachments folder? I found that particularly odd and suggested perhaps it could use little more fine tuning.
The first link (JStage) also gives public access to the PDF, but the article.pdf is one click away after the Refs»Pubmed»Full text(if available) maneuver. Again I wonder if bookends can open the full text page without a problem and the PDF link is right on that page, how come it couldn't download it.
Perhaps you have perfectly good reasons for such behaviour and as a mere end user I may be just oblivious to them – but I thought I would give my 2¢ with the hope it'd be useful.
What seems simple to you is not necessarily so. Each of the examples you sent would require a custom search and parser. Can it be done -- of course. But that is NOT "fine tuning". That's brute force. Multiply it by several thousand and it should be obvious it's not a good idea. As I said, we cover most of the major publishers and journals (and a lot of very minor ones). If there are particularly important ones (by that, I mean of general interest, not subspecialty journals) that aren't included, I'm open to adding them if it's possible.
Jon
Sonny Software
Jon
Sonny Software