On Feb 24, 2005, at 8:55 AM, Dr Robert Sanderson wrote:
Web searches are looking for data that you do not know. One presumes that the opposite is true when you are searching within your own database of citations.
I don't think this is necessarily the case. Often I do know generally what I need (mid-1990s article by author x and y), but equally commonly, I'm looking for "anything related to topics a and b", where "anything" could include annotation content.
Have you read Tim Bray's discussions on search interfaces? Research and everyday anecdote shows that when presented a choice between using a simple query interface, and a complex one, users overwhelmingly choose the former. The reason is because it's a PITA to have to switch to another interface and work across multiple fields. I certainly don't want to do that when I'm writing just to find some references.
In my eXist-based database, I can easily and quickly find whatever I need (including as I said, stuff that I actually do need but completely forgot about) simply by using a google query syntax.
The underlying technology is not just dumb, though; it using some sort of proximity matching a la Lucienne to rank results.
Demo here, though I'm not sure how many records are loaded (it's bundled with eXist anyway, so you can always try for yourself):
http://demo.exist-db.org/exist/mods/
Bruce
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
