History
Sphinx development was started back in 2001, because I didn't manage to find an acceptable search solution (for a database driven Web site) which would meet my requirements. Actually, each and every important aspect was a problem:
-
search quality (ie. good relevance)
-
statistical ranking methods performed rather bad, especially on large collections of small documents (forums, blogs, etc)
-
search speed
-
especially if searching for phrases which contain stopwords, as in “to be or not to be”
-
moderate disk and CPU requirements when indexing
-
important in shared hosting environment, not to mention the indexing speed.
Despite the amount of time passed and numerous improvements made in the other solutions, there's still no solution which I personally would be eager to migrate to.
Considering that and a lot of positive feedback received from Sphinx users during last years, the obvious decision is to continue developing Sphinx (and, eventually, to take over the world).
