A few people have emailed arguing that I have Google Scholar wrong. After all, they say, I have sung the praises of narrow search a number of times in the past, so aren’t I being a tad inconsistent?
Not at all. I still love narrow search, and I think the biggest threat to Google comes from highly domain-specific search, companies that really know a particular area (like company management structure, or careers, or SEC filings) both in depth/context and longitudinally, and that can charge a premium to people who need that data organized with a rich ontology/taxonomy and presented in highly manipulable fashion. No question.
Is it there in Google Scholar? I don’t think so (yet), but I agree it is the right direction. Most of what is interesting in search going forward centers on constraining the relevant domain, so we will see more, not less, of this from Google (and its ilk). The days of throwing high-horsepower spiders at an over-broad page universe in hopes of accumulating higher index counts than some competitor is like a replay of the muscle-car wars of the 1970s — and they are already ending.