Nasty Dose Of Reality 3
If it’s not recall and precision, what’s the biggest concern?
Throughput.
Google handles 100 million queries /day, most in < 0.5 sec.
(2,000 queries / sec. peak).
Because the Web is so big, any scheme involving complicated
queries and/or document analysis is just not feasible.
Trying to work within the existing search engine framework is
probably a bad idea.