Thanks to everyone for all the useful suggestions. I tried out the zotero firefox add-on, and I think that it's really quite impressive. It's not perfect (I'll elaborate below) but it goes a very long way toward the features I want and even adds some nice things I hadn't thought of. It's not perfect, but I think I'm going to use it very happily and I'd suggest others take a look.
Zotero does a lot of what I wanted. It has some schemes to parse the web pages of various journals and database sites, so that all you have to do it hit an icon in the address bar and it grabs all the bibliographic information for the article. It can be set to grab a copy of HTML articles and/or download the PDF of the article. All this information is stored under an entry for that item in the database, and it allows you to tag the entries. It will allow you to search by tags or by full text search (to get that to work for the PDFs takes a bit more work, but it was still very easy in my case). This really satisfies most of what I wanted and it incredibly easy to use, with grabbing most papers taking only one click. It offers a number of other useful features, some that I wanted but didn't mention and some I hadn't really considered. For each entry in the database you can attach notes or attach other files of your choosing. Also, if you take a snapshot of the HTML version of an article, Zotero will allow you to highlight and annotate your snapshot of the article, in addition to any stand-alone notes you might add to the database entry. You can also make collections inside you "library" for organizing the contents, where an entry can appear in multiple collections. Overall, I think Zotero is a very impressive piece of software, but there are some caveats. The one feature I really want that it doesn't directly provide is a facility to track read and unread papers. You can use the tagging for this purpose, but the drawback is that if you export entries to, say, BibTeX then these tags will ago along for the ride, which is probably undesirable. The other big drawback I've discovered is that the export to BibTeX has some issues, including problems with conversion of tags to keywords and the handling of special characters. Importing from BibTeX has similar problems. Still, I think that dealing with these issues won't be a lot more work than I'm doing now to get bibliographic info into and out of my BibTeX database, and it's a good trade-off for the many other advantages of Zotero. Another issue to be worked out is synchronization between computers. I like to go between my work and home desktop computers and sync my data (using Unison) so that I can just keep on working when I get to the other computer as though I'd never moved. It looks like it may not be that easy to make this work smoothly with Zotero, but it looks like there are some ways to attack the problem and they're working on this for their next release. The final concern I have is about being able to get my data back out of Zotero if I later want to move to some other system. I don't think this is such a big concern because Zotero is open source, seems to use a standard SQLlite database, and keeps all the copies of papers in a plain old directory structure. Hope that's helpful to others. Regards, Nick Gaurav Aggarwal wrote: > You may find this useful. Its called Zotero and is a firefox plugin > > http://www.downloadsquad.com/2006/11/27/zotero-firefox-bibliography-plugin/ > > -gaurav > > > On Wed, Oct 1, 2008 at 2:08 PM, Nick Cummings <[EMAIL PROTECTED] > <mailto:[EMAIL PROTECTED]>> wrote: > > For a while I've really wanted to have some sort of database that > would > allow me to keep track of papers relevant to my research. Basically, > right now I just download papers I find that are of interest, and put > them in a hierarchy of directories by subject. This has a couple of > problems: 1) It takes time to do the organizing. 2) The subject > matter > of papers is not hierarchical, it's more of a web. 3) There's no easy > way to store metadata about the paper (e.g., author & title) such that > it can be browsed through without opening every paper. 4) No > convenient > way to track read/unread status. 5) It's not searchable. I'm hoping > there's a piece of software one can use on Linux to help with this > sort > of thing, either software specifically for this purpose or one that > could be adapted without very much work (or much programming > expertise). Surely some of you bright academics have tried to find a > solution to this sort of problem before. :-) I'm also open to web > services for these purposes, my only hesitation is that I may then end > up with my info locked into it with no possibility to migrate later. > > I've come up with some idea for solutions, but none seem all that > good: > 1) I could simply keep a spreadsheet (or text file) with each paper's > associated information (bibliographic info, read/unread status, > tags/keywords) and the file name of the paper (if it exists). 2) > Try to > use my BibTeX bibliographic database and Pybliographic GUI to keep > track > of all the info about papers and, again, point to file locations (not > sure how feasible this is). 3) Hope I can find a URL for the > abstract of > each paper and use del.icio.us <http://del.icio.us> to bookmark > and tag each one. Make sure > to put all bibliographic info in the notes section. > > Of those, #1 seems like the most plausible solution but far from > ideal. > > Regards, > > Nick > >
signature.asc
Description: OpenPGP digital signature
