| Home | Geschichten | Kunst | Computer | Tindertraum |
A Bloom filter encoding someone's contact information can be checked to see whether it contains a given name or email address, but it can't be coerced into revealing the full list of keys that were used to build it.
Be sure to read all of it. This 'check only, no list' feature and the extreme compactness memory wise seem like very handy tricks.
Bloom Filters for finding annotations.
Several Bloom filters can be ored together offset by various small numbers of bits. Now hits can be related to a particular original filter. If you think this thru you will find that such a merged filter is very efficient to read. Of course the merged filter may be originally constructed as merged.
And here is an alternative approach:
FS Consulting Inc's Weblog
Personally if I need a structure to store lots of keys and have a very fast lookup, I use tries. There is a nice diagram here and some definition and source code here. There are a few nice things about tries. They load pretty fast, they are fast assuming you have loaded them in a balanced fashion, you can mark data as active or inactive (though I suspect that you can also delete data) and lookup is thread-safe (assuming you have coded it that way.)
[ by Martin>] [permalink] [similar entries]
similar entries (vs):
similar entries (cg):
no similar entries (yet?)
Martin Spernau
© 1994-2003
Big things to come (TM) 30th Dez 2002
(Organic) machinery
Oblique Strategies,
Ed.3
Brian Eno and Peter Schmidt
amazon.de Wunschliste
usefull links:
Google Graph browser
Traumwind 6-Colormatch
UAV News
powered by SBELT