Google Schoolar + identify and extract citations

khazaei · March 4, 2010

Hi
There are two useful features in Mnedeley (http://www.mendeley.com) which I can not find them in Zotero.
Maybe they are already included in Zotero and I could not find them.

1-In Mendeley, if you manual add a citation to your library by writing just the title of the article (and/or authors), you can validate it and get more information about it an option that search the new reference in Google-scholar and extrac all its necessary information (journal, year, DOI, web address and etc.) automatically.
Is there such an option in Zoero?

2-Mendely has the ability to identify all citations corresponding to a given file (PDF, word, RTF...) and extract them as a new collection in your library. Is it also possible in Zotero?

adamsmith · March 4, 2010

no, neither of those exist in Zotero. I don't quite follow what 2) does - are those citations from google scholar?

ajlyon · March 4, 2010

I feel like (1) could build off of the current system for fetching metadata for PDFs-- an option to right-click on any bibliographic entry to send the existing data to a bibliographic database as a look-up, then pulling up the current partial data and the retrieved data in the comparison/merging interface that is currently used for resolving sync conflicts. Or, if there are multiple credible results, it would first show the select a match screen that you get after fetching metadata for a PDF, then a merge screen from the current sync system.

In fact, couldn't we use the sync conflicts resolution system for duplicate item issues as well?

sean · March 4, 2010

Given that users can achieve request #1 merely by navigating to Google Scholar and entering a query there, it seems like a lot of coding for fairly minimal streamlining. That said, ajlyon's proposal to use the existing conflict resolution dialog to compare and resolve metadata is an excellent one. Request #2 frankly doesn't work that well for a variety of technical reasons. We've been looking into automatic citation extraction since 2006, and the problem is simply that more noise than signal ends up getting introduced, particularly once users branch out into formatting styles beyond the sciences. That said, there's no reason why a developer couldn't add a plugin using c2bib to achieve a similar result. It would be a fairly straightforward project, and likely uses the same approach (regular expressions) adopted by Mendeley.