Full-text search package?
Alan.Kay at squeakland.org
Tue Jan 22 19:33:28 UTC 2002
Hey Scott --
Do it, and it will enter the holy image ....
At 2:12 PM -0500 1/22/02, Scott A Crosby wrote:
>On 22 Jan 2002, Cees de Groot wrote:
>> Now, before I go and hack something simplistic (I'm thinking about a
>> simple word index with frequencies, and being too lazy for a real
>> persistence engine and recognizing that reiserfs really is a database,
>> it'll be spread out over an awful lot of files ;-)), is there a decent
>> text indexing/searching package available for Smalltalk or Squeak?
>Hack'er one up yourself? Reverse indexes aren't all that slow or large.
>Roughly comparable to the size of the origional data.
>A reverse index is for each word, you link to a list (or rather, a set)
>of things that include that word. Then do set union-disjunction operations
>to implement boolean matching, prefix-matching, and substring matching.
>I love languages with featureful internal libraries, I could probably code
>up the engine myself in an afternoon.
>Or, would you rather do it? (whats your luck in getting your stuff into
>the image when you've completed it?)
More information about the Squeak-dev