As described by Walt Crawford in Losing What Counts: The Swamping Phenomenon, swamping happens when "the stuff that counts" disappears "under a flood of other vaguely similar stuff", such as what happened when Amazon introduced "Search in the Book". Though passages in books suddenly became findable, simply finding a book when you knew the title was problematic. Swamping can arise with any search engine when the amount of full-text content grows dramatically, especially when full-text resources are mixed with bibliographic ones. Luckily, Amazon was able to tweak their relevancy ranking to overcome the swamping phenomena.
Walt concludes that "The solution isn't to avoid large databases, heterogeneous databases, or searchable full text. Sensible defaults, fielding where feasible, good advice and thoughful results handling can keep swamping under control." Google has to fight the effects of swamping every day on a huge scale, but swamping can happen to the best of us, even when working with databases on smaller scale.
Leave a comment