[tlug] Searching for kanji strings: indexing by FreeWAIS for speed

[Jim <jep200404@example.com>]

"Stephen J. Turnbull" wrote:

> FreeWAIS sounds like the thing.  It understands proximity searches
> (ie, two strings X words apart).  If you tell it that every UTF-8
> character is a word ... but you'll end up with indexes 10-20X as big
> as your text, I'm afraid.

My sister bought a 300GB drive for $70. 
For text that is hundreds of Megabytes, 
a 10 to 20 times bigger index should be OK as long as it is fast.