This post from Dig_Ref announces the IIS’ launch of Reference Extract:

Reference Extract is a targeted web search engine. It is built from the expertise of over 60 AskA services geared to the education audience. The Virtual Reference Desk team has identified high-quality archives of FAQ’s and previously asked questions. These sites were then indexed, and the result is an easy to use, quality oriented search engine: Reference Extract.

Some of the subsequent posts to Dig_Ref say things like, there are poor quality sites in there, or such-and-such a search came back with irrelevant hits. I say, well sure, that’s bound to happen. After all, the URLs in Reference Extract were included in dig ref answers, in response to specific questions, & now they’re being mined for answers to different questions. I’d expect a disconnect. I’ve written that dig ref transactions, because they can be captured, may be used as annotations to information resources that they refer to. In other words, the resources provided in response to a question on subject X are provided as good resources on subject X. (Under what circumstances would a reference librarian ever provide a resource that they know to be poor?) So the question is, can, & how can those same resources be mined for answers to different questions than they were originally used to answer?

I haven’t yet spoken to the folks at the Institute about this, but I will when I get some time, & I’ll report back here. Anyway, I think this is potentially a brilliant idea & I wish I’d thought of it.

Dave has also started a Reference EXTRACT blog.