home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.theory:2837 sci.crypt:6557 sci.math:17866 rec.puzzles:8295
- Path: sparky!uunet!usna!math3!mdm
- From: mdm@math3.sma.usna.navy.mil (Mark D. Meyerson -- math FACULTY <mdm@sma.usna.navy.mil>)
- Newsgroups: comp.theory,sci.crypt,sci.math,rec.puzzles
- Subject: Re: Looking for random permutation generation algorithms
- Message-ID: <2614@usna.NAVY.MIL>
- Date: 8 Jan 93 21:33:46 GMT
- References: <1993Jan6.014749.15323@ee.ubc.ca> <1993Jan7.165939.11149@cs.cornell.edu> <2607@usna.NAVY.MIL> <1993Jan8.163909.20820@cs.cornell.edu>
- Sender: news@usna.NAVY.MIL
- Reply-To: mdm@math3.sma.usna.navy.mil (Mark D. Meyerson -- math FACULTY <mdm@sma.usna.navy.mil>)
- Followup-To: comp.theory
- Lines: 13
-
- In article <1993Jan8.163909.20820@cs.cornell.edu>, karr@cs.cornell.edu
- (David Karr) writes:
- |> This is really the major omission from the algorithm. I had meant
- |> to compute both k mod b *and* k div b. k mod b is used to index the
- |> array for the current selection; k div b is then copied to k for the
- |> next round.
- |>
-
- I agree now. This seems to be equivalent to the
- straightforward but perhaps harder to code recursive
- method. I've coded and checked Karr's algorithm
- for N=1 to 6.
- Mark D. Meyerson
-