On Wed, Jan 5, 2011 at 3:50 PM, Brian Brian.Mingus@colorado.edu wrote:
On Wed, Jan 5, 2011 at 3:44 PM, David Gerard dgerard@gmail.com wrote:
On 5 January 2011 22:40, geni geniice@gmail.com wrote:
Basically no If you look at even [[Template:Cite web]] it requires stuff that you have to go hunting for (author). You could construct something for popular websites (BBC say) which have a standard format.
Sounds like something we could add really quite a lot of special cases to. I wonder how many we would need to have decent coverage in practice. Has anyone done a survey of what sources we actually use in references? The long tail will be *huge*, but does the en:wp community have any favourites?
- d.
I have created a tool called WikiPapers that my lab has used for several years that does something similar to this. It is designed around scientific papers. It allows you to highlight the title of an article on any web page and then click it a bookmarklet and it will use various APIs on the web to get the associated metadata and add it to your wiki. It can optionally pass the URL to one of many URL scrapers such as Connotea and CiteULike. I am currently refactoring the code for use in a new project called WikiScholar. The old code supports PubMed, Google Scholar, Connotea and CiteULike, whereas the new code only supports PubMed right now. The new code, however, makes it much simpler to add new importers with its class-based infrastructure.
If anyone is interested in this project and can code in Python or PHP please let me know. I am actively developing it now. I'm interested in folks who would like to dedicate some time to writing importers for specific APIs.
Cheers, Brian
PS: The Google Code url is: http://code.google.com/p/wikipapers/