From: Adam R. M. <ama...@ma...> - 2007-03-30 03:58:14
|
On Mar 29, 2007, at 19:08, Michael McCracken wrote: > Adam, I may have misled you about the PMID file naming - I poked > around PubMed today and only got useless non-unique file names (like > "fulltext.pdf" - gee, thanks!) > I think that I must have been remembering an app that saved files > gotten from search results that way... Sorry. Maybe hubmed does that? I've heard the same thing, but I generally =20 use sciencedirect or lanl. > Users: was I wrong - is it more common than I think to download files > named by their PMID? > > > =46rom their website, Bookends finds a DOI somewhere in the file: > >> =95 Import a reference from PubMed automatically when attaching its = pdf >> >> If an attached pdf includes a doi (Digital Object Identifier) that >> Bookends can resolve and that is listed on PubMed, Bookends will >> offer to download the reference information and attach the pdf to >> the new reference. This will work even if you don't have access >> privileges to the pdf online. Mac OS X 10.4 (Tiger) or later is >> required. > The movies for Papers show them doing a nifty version of copy and > paste to build a search from the contents of a PDF, but they don't > seem to do anything automatically. I haven't looked at those; I've just been hearing how great Papers is =20= at automatically doing everything :). In the next version of BibDesk =20= it will be pretty easy to open a search group, paste in the pmid, and =20= get the results. > =46rom a quick scan of some papers that do have a DOI (it's in the > MEDLINE export data), > the best we could do is search the PDF for the string "DOI" and grab > what's after it. That's sort of lame, but it's doable; we already do something similar =20= for using PDF metadata (hidden pref). I guess a regex for PMID would =20= also be possible (but lame as well). The problem with a doi is that =20 it's not required to have a "DOI" string anywhere, so that's not very =20= reliable either. adam |