On Sat, Aug 05, 2000 at 07:00:44AM +0300, Henry Hemming wrote: > > You really need to read a data structures book before arguing with > > everyone in a public forum like this. For a start, start reading the > > balanced trees chapter of Knuths "Searching and Sorting" and keep on > > reading past that into the multi-way tree section. > Well, you shouln\t even be on this mailing list if you havent read it, or > smth to compare it with. After all, freenet is all about searching and > sorting. Also some genetic algorithms and neural network literature should't > hurt. Not to mention cryptography.
I looked up on hashes, and what Scott said about hashes was correct. Anyways, I figured out that a 65536 entry hash (this length isn't arbitrary) would be a good idea because it greatly reduces the chances of collisions and allows you to use a hash function which treats a key as 10 * 16 bits and simply XORs each 16 bits (and does not require a modulus). On a 32 bit machine, the hash would be 256K. Note that this hash would not be stored on disk (even though it would be theoretically possible), but would be constructed at startup (which would normally be during a computer's boot phase) and would be updated when requests for unknown files were successful. Anyways, I don't have a copy of any book by Knuth. The only book on algorithms that I have is "Mastering Algorithms with C" by Kyle Loudon (which is one of the many O'Reilly books). -- Travis Bemann Sendmail is still screwed up on my box. My email address is really bemann at execpc.com. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 1650 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20000805/3c4b2f57/attachment.pgp>
