There seems to be a rapidly growing interest in expanding the range of 
empirical data sources for testing linguistic hypotheses, with books and 
special issues and conferences in this general area becoming more and more 
common. For a conference that we recently held ourselves here in Taiwan, I 
threw together an arbitrary list of references and links: 
 
http://www.ccunix.ccu.edu.tw/~lngproc/IWGElinkse.htm 
 
But it seems to me that there should be a more ambitious, well-organized 
attempt to provide a Web-based portal to this kind of information. This 
might include a searchable paper archive, perhaps using the Ashing software: 
 
http://software.auf.net/ashing/ 
 
... and/or a wiki, with advice on experimental design and statistics, links 
to corpora and software, and suchlike stuff. 
 
I was thinking about starting such a resource myself, but I'm not sure our 
server could handle the traffic, and anyway this kind of thing requires an 
active community (not to mention people who actually know how to write and 
maintain server-side software). 
 
The main difficulty with such a project would be drawing the line between 
''empirical'' linguistics and ''ordinary'' linguistics, since of course no 
sharp line exists. But that's what discussions are for.