Academia Stack Exchange is a question and answer site for academics and those enrolled in higher education. It only takes a minute to sign up.
Sign up to join this community
Mendeley is (mainly) a proprietary social network to share basic citation data and research papers. see also
Are there any free and open source substitutes for this service?
A technical possibility: users could collect BibTeX data and a hash and share this information. It would be very useful to have such a free service because every journal provides the citation data in a different format. That includes false field entries, broken files and hidden download buttons on the website.
The only alternative that comes to my mind is Zotero:
The Zotero standalone client is cross-platform and open-source (AGPL licence), and it can be run on its own or synchronized with the web version. The web service is free to use up to a fixed storage quota, with paid storage available, and there is an open-source implementation of the dataserver available if you want to roll your own. The local client stores its data in SQLite format so in principle your data is not locked in, but the database is relatively hard to trawl externally; however, since the client is open source there are relatively few future-proofing concerns.
Haven't tried it yet but https://www.docear.org/ seems to stand out as one of the leading open-source alternative to mendeley.
Docear is basically a marriage of JabRef and Freeplane. It uses JabRef as a backend for its reference management and Freeplane to organize references, annotations you make in the pdf, and any other information (including images, links and crossreferences) in a mind-map.
As the original author of this answer pointed out it is indeed open-source—licensed under the GNU General Public License.
I really liked Mendeley's potential but got frustrated with both their pricing model (maybe I just never learned how to use the software correctly) and it consistently butchering imported BibTeX entries.
I've been a pretty happy BibDesk user for a long time, it is true open source software, but unfortunately it has not been ported outside of the OS X environment, so this is only a qualified answer.
I have tried any number of reference management programs (from EndNote to Mendeley and back), and have come to the conclusion that the most important criterion is that the reference manager directly, natively writes in .bib, or some other text-based, open-source format.
This has several advantages:
This leaves a pretty small number of editors:
We have developed an open source solution, downloadable from GitHub at https://github.com/scientilla/scientilla, that allows users to collaboratively share and refine their scientific bibliographic metadata.
The system relies on a "peer-to-peer" and "open-data" approach as well as on a "clone-and-refine" algorithm. It can import data from external web services.
The more the system is used, the more the information that flows on the network become clean.
Moreover, using Scientilla, any user can obtain the whole metadata shared through the network.
Finally I found I, librarian on http://i-librarian.net/ which is a kind of mixture between Filestorage-Server, JabRef and a personal open source Mendeley server.
Interested users may try the demo account on the website.
In creating such a system, the cost of the software and the expense of running the server(s) to host the data is microscopic compared to the expense of the army of knowledgeable people checking out references, cleaning up the barely legal (and semantically often complete nonsense) BibTeX entries offered by some journals, knit them up from scratch where not even the above is available.
Better create your central database, share it with your research group. Have everybody chip in by adding references to interesting papers found while browsing, whatever they cite in their publications/theses, and perhaps keep another file with publications generated locally (comes handy when asked for "publications of the group last year" or so). Publish guidelines, enforce them, perhaps occasionally organize a refer-thon to clean up entries and fill out missing details. If you ask everybody who reads a paper/document to add a short (3-5 lines) summary/abstract (they'll have to write one anyway for "state of the art"), a URL or other pointer where to find the document, you create an (at least locally) very valuable resource.
If it grows enough, or as a side product of the work, publish it on the webpage of the group. Might even ask for contributions from outside while you are at it.
No, there isn't anything that is an exact copy of Mendeley, the server, the clients, the API, and the content, but all composed of 100% open source software.
For the record, Mendeley does contain some open source and the API is freely available, so if you wanted to build an application of service to the scholarly community, you'd probably be better off building your own open source client but leveraging the messy and difficult bits that we've already solved and make available via API, such as the metadata extraction and impact tracking bits.
