Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.c > #83511 > unrolled thread

Testing nodes and lists for hashtable

Started byAlla _ <modelling.data@gmail.com>
First post2016-03-09 00:52 -0800
Last post2016-03-30 07:00 -0700
Articles 20 on this page of 330 — 24 participants

Back to article view | Back to comp.lang.c


Contents

  Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 00:52 -0800
    Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-09 10:13 +0000
      Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 02:19 -0800
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 02:23 -0800
        Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-09 13:00 +0000
    Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-09 02:38 -0800
      Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 02:59 -0800
        Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-09 10:35 -0800
          Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-10 00:20 -0800
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 02:56 -0800
      Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-09 13:11 +0000
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 06:37 -0800
          Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-09 15:50 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 08:19 -0800
              Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-09 17:18 +0000
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-10 00:11 -0800
                  Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-10 08:48 +0000
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-10 04:17 -0800
                  Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-10 11:12 +0000
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-10 04:16 -0800
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-10 04:20 -0800
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 08:41 -0800
              Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-09 17:27 +0000
                Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-09 22:41 -0600
                  Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-10 10:38 +0000
                    Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-10 09:37 -0600
                      Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-10 15:51 +0000
                        Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-10 18:42 -0600
                        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 04:17 -0800
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-10 00:18 -0800
                  Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-10 08:51 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 08:50 -0800
    Re: Testing nodes and lists for hashtable jadill33@gmail.com - 2016-03-09 07:16 -0800
    Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-09 09:27 -0600
      Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-09 07:46 -0800
        Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-09 10:56 -0800
          Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-10 00:25 -0800
            Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-10 08:55 -0800
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 01:14 -0800
      Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-11 01:53 -0800
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 02:25 -0800
          Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-11 12:31 -0800
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-12 00:03 -0800
              Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-12 09:04 -0800
        Re: Testing nodes and lists for hashtable mark.bluemel@gmail.com - 2016-03-11 02:30 -0800
          Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 02:39 -0800
            Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-11 02:52 -0800
              Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 04:09 -0800
              Re: Testing nodes and lists for hashtable mark.bluemel@gmail.com - 2016-03-14 08:41 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 02:34 -0800
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 02:41 -0800
          Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-11 12:14 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 04:20 -0800
              Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-11 12:33 +0000
              Re: Testing nodes and lists for hashtable Randy Howard <rhoward.mx@EverybodyUsesIt.com> - 2016-03-11 12:12 -0600
                Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-11 10:24 -0800
                Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-11 19:04 +0000
                  Re: Testing nodes and lists for hashtable Randy Howard <rhoward.mx@EverybodyUsesIt.com> - 2016-03-11 14:20 -0600
                  Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 23:55 -0800
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 04:23 -0800
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 04:25 -0800
              Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-11 12:36 +0000
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 05:56 -0800
                  Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-11 14:55 +0000
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 07:22 -0800
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 07:29 -0800
                      Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-11 16:01 +0000
                        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 23:41 -0800
                          Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-13 13:48 +0000
                    Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-11 20:20 +0000
                      Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-13 13:06 +0000
                  Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-12 17:41 -0600
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-13 05:03 -0700
                      Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-13 13:50 +0000
                        Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-13 10:32 -0500
                          Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-13 09:18 -0700
                          Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-13 16:31 +0000
                            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-13 10:17 -0700
                      Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-13 10:17 -0500
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-13 09:06 -0700
                      Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-13 09:51 -0700
                        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-13 10:31 -0700
                      Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-13 12:41 -0500
      Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-11 08:41 -0800
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-11 23:54 -0800
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-12 08:10 -0800
      Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-12 08:17 -0800
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-12 08:19 -0800
          Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-12 09:16 -0800
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-13 04:17 -0700
          Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-12 19:37 +0000
            Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-13 11:20 -0500
              Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-13 21:37 +0000
        Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-12 09:48 -0800
          Re: Testing nodes and lists for hashtable raltbos@xs4all.nl (Richard Bos) - 2016-03-12 20:31 +0000
          Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-12 21:18 +0000
          Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-12 17:54 -0600
          Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-13 04:25 -0700
            Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-13 04:44 -0700
              Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-13 12:31 -0500
                Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-13 11:14 -0700
                  Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-13 16:17 -0500
                    Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-13 14:37 -0700
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-13 11:43 -0700
      Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-13 12:08 -0700
      Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-13 16:31 -0500
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 01:12 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 01:21 -0700
          Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-14 13:48 -0500
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 01:43 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 01:35 -0700
          Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-14 13:44 -0500
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 01:39 -0700
      Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-13 23:04 +0000
        Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-13 16:09 -0700
          Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-13 17:54 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 01:58 -0700
          Re: Testing nodes and lists for hashtable fir <profesor.fir@gmail.com> - 2016-03-14 03:08 -0700
          Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-14 10:59 +0000
            Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-14 04:59 -0700
              Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-14 13:15 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 11:09 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 03:18 -0700
          Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-14 11:03 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 11:17 -0700
              Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-14 15:16 -0700
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 01:59 -0700
                  Re: Testing nodes and lists for hashtable Barry Schwarz <schwarz45@yahoo.com> - 2016-03-15 20:13 -0700
                    Re: Testing nodes and lists for hashtable supercat@casperkitty.com - 2016-03-15 22:09 -0700
                    Re: Testing nodes and lists for hashtable BartC <bc@freeuk.com> - 2016-03-16 11:53 +0000
                    Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-16 09:39 -0500
                      Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-16 07:57 -0700
                        Re: Testing nodes and lists for hashtable supercat@casperkitty.com - 2016-03-16 08:38 -0700
                          Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-16 09:04 -0700
                            Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-16 11:22 -0500
                            Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-16 10:42 -0700
                              Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-16 10:57 -0700
                                Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-16 11:57 -0700
                        Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-16 10:43 -0500
                      Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-16 10:59 -0700
                        Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-16 13:22 -0500
                          Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-16 16:09 -0700
                            Re: Testing nodes and lists for hashtable supercat@casperkitty.com - 2016-03-16 17:03 -0700
                            Re: Testing nodes and lists for hashtable raltbos@xs4all.nl (Richard Bos) - 2016-03-20 13:04 +0000
                            Re: Testing nodes and lists for hashtable Tim Rentsch <txr@alumni.caltech.edu> - 2016-03-30 09:07 -0700
                              Re: Testing nodes and lists for hashtable luser droog <luser.droog@gmail.com> - 2016-03-31 20:34 -0700
          Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-14 14:04 +0000
          Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-14 11:07 -0700
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 03:49 -0700
      Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-14 13:03 +0000
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 11:24 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-14 11:41 -0700
          Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-14 19:35 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 01:53 -0700
              Re: Testing nodes and lists for hashtable luser droog <luser.droog@gmail.com> - 2016-03-15 02:03 -0700
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 02:09 -0700
                  Re: Testing nodes and lists for hashtable luser droog <luser.droog@gmail.com> - 2016-03-15 02:13 -0700
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 02:22 -0700
                  Re: Testing nodes and lists for hashtable luser droog <luser.droog@gmail.com> - 2016-03-15 02:15 -0700
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 02:28 -0700
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 02:30 -0700
              Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-15 09:13 +0000
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 02:28 -0700
                Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-15 03:15 -0700
                  Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-15 13:26 +0000
                  Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-15 09:18 -0700
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 03:19 -0700
                  Re: Testing nodes and lists for hashtable David Brown <david.brown@hesbynett.no> - 2016-03-15 15:21 +0100
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 08:39 -0700
                      Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-15 09:26 -0700
                        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 00:10 -0700
                  Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-15 09:19 -0700
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 00:08 -0700
                      Re: Testing nodes and lists for hashtable David Brown <david.brown@hesbynett.no> - 2016-03-16 08:55 +0100
                      Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-16 09:17 -0700
                        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 09:25 -0700
          Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-14 13:06 -0700
            Re: Testing nodes and lists for hashtable Gareth Owen <gwowen@gmail.com> - 2016-03-14 20:29 +0000
            Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-14 21:03 +0000
              Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-14 14:57 -0700
                Re: Testing nodes and lists for hashtable raltbos@xs4all.nl (Richard Bos) - 2016-03-14 22:35 +0000
              Re: Testing nodes and lists for hashtable David Brown <david.brown@hesbynett.no> - 2016-03-16 19:47 +0100
            Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-14 14:26 -0700
      Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-14 11:37 -0700
      Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-14 11:37 -0700
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-15 08:58 -0700
      Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-15 16:12 +0000
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 00:04 -0700
      Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-15 10:48 -0700
        Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-15 18:00 +0000
          Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-15 19:20 +0000
            Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-15 21:24 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 00:50 -0700
              Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-16 11:19 +0000
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 04:37 -0700
                  Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 11:51 +0000
                    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 05:20 -0700
                      Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 07:13 -0700
                        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 09:05 -0700
                Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 11:45 +0000
                  Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-16 13:46 +0000
                    Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 14:00 +0000
                    Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-16 07:05 -0700
                      Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-16 16:42 +0000
                        Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-16 10:21 -0700
                          Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 17:33 +0000
                            Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-16 10:44 -0700
                              Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 18:00 +0000
                                Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-16 11:08 -0700
                                  Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 18:19 +0000
                                    Re: Testing nodes and lists for hashtable Gareth Owen <gwowen@gmail.com> - 2016-03-16 19:38 +0000
                                      Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-16 14:00 -0700
                                      Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 21:52 +0000
                                        Re: Testing nodes and lists for hashtable Gareth Owen <gwowen@gmail.com> - 2016-03-16 22:12 +0000
                                Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-16 13:39 -0500
                          Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-16 15:39 -0700
                            Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-17 02:59 +0000
                              Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-17 04:20 -0700
                                Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-17 23:46 +0000
                                  Re: Testing nodes and lists for hashtable Jerry Stuckle <jstucklex@attglobal.net> - 2016-03-17 20:04 -0400
                                    Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-18 01:05 +0000
                                      Re: Testing nodes and lists for hashtable Jerry Stuckle <jstucklex@attglobal.net> - 2016-03-17 22:35 -0400
                                  Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-17 17:19 -0700
                                    Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-18 01:39 +0000
                                      Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-18 02:54 -0700
                                        Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-18 09:08 -0500
                                          Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-18 11:20 -0700
                                        Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-18 21:29 +0000
                                          Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-18 17:48 -0700
                            Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-16 22:09 -0500
                              Re: Testing nodes and lists for hashtable raltbos@xs4all.nl (Richard Bos) - 2016-03-17 11:49 +0000
                                Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-17 08:45 -0700
                                Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-17 12:11 -0500
                                  Re: Testing nodes and lists for hashtable supercat@casperkitty.com - 2016-03-17 10:31 -0700
                                    Re: Testing nodes and lists for hashtable Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-03-17 11:09 -0700
                                      Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-17 16:15 -0500
                                      Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-18 00:39 +0000
                      Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-16 12:04 -0500
                    Re: Testing nodes and lists for hashtable Steve Thompson <stevet810@gmail.com> - 2016-03-17 14:21 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 00:53 -0700
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 01:00 -0700
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 01:24 -0700
              Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-16 11:08 +0000
                Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 11:25 +0000
          Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 00:41 -0700
            Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-16 08:27 -0500
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 06:43 -0700
      Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 14:07 +0000
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 08:16 -0700
          Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-16 15:25 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 08:30 -0700
      Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-16 20:32 +0000
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-17 03:26 -0700
          Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-17 16:15 +0000
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-17 11:07 -0700
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 08:47 -0700
      Re: Testing nodes and lists for hashtable Stephen Sprunk <stephen@sprunk.org> - 2016-03-16 10:54 -0500
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 09:35 -0700
      Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-16 11:31 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-17 02:53 -0700
          Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-28 18:08 -0700
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-30 02:07 -0700
              Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-30 09:45 -0700
                Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-30 17:54 +0100
      Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-16 20:26 +0000
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-17 03:00 -0700
          Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-17 16:05 +0000
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-16 09:10 -0700
      Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-16 12:24 -0700
        Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-16 13:16 -0700
          Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-16 16:20 -0700
            Re: Testing nodes and lists for hashtable supercat@casperkitty.com - 2016-03-16 16:54 -0700
            Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-16 17:27 -0700
        Re: Testing nodes and lists for hashtable supercat@casperkitty.com - 2016-03-16 13:18 -0700
      Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-16 20:21 +0000
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-17 04:23 -0700
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-17 04:47 -0700
      Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-17 05:40 -0700
        Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-17 12:49 +0000
          Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-17 06:45 -0700
        Re: Testing nodes and lists for hashtable mark.bluemel@gmail.com - 2016-03-17 10:00 -0700
          Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-17 12:40 -0700
    Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-28 12:00 -0700
      Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-28 12:37 -0700
        Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-28 17:43 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-30 01:31 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-30 01:55 -0700
          Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-30 10:23 +0100
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-30 02:32 -0700
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-30 02:33 -0700
          Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 07:04 -0700
            Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-30 15:17 +0100
              Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 07:46 -0700
                Re: Testing nodes and lists for hashtable Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-03-30 16:41 +0100
                  Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 09:16 -0700
                    Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 09:18 -0700
                    Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 09:37 -0700
          Re: Testing nodes and lists for hashtable Barry Schwarz <schwarzb@dqel.com> - 2016-03-30 10:32 -0700
            Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-30 12:10 -0700
      Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-28 17:42 -0700
        Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-29 08:51 -0700
          Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-29 09:56 -0700
            Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-29 12:09 -0700
              Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-29 16:34 -0700
                Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-29 16:50 -0700
                  Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-29 17:17 -0700
                  Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-30 02:30 -0700
                Re: Testing nodes and lists for hashtable Richard Heathfield <rjh@cpax.org.uk> - 2016-03-30 00:52 +0100
                  Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-29 17:27 -0700
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-30 02:26 -0700
                  Re: Testing nodes and lists for hashtable mark.bluemel@gmail.com - 2016-03-30 07:09 -0700
                  Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 07:33 -0700
                    Re: Testing nodes and lists for hashtable Jerry Stuckle <jstucklex@attglobal.net> - 2016-03-30 11:32 -0400
                      Re: Testing nodes and lists for hashtable mrs@kithrup.com (Mike Stump) - 2016-04-29 08:21 +0000
                        Re: Testing nodes and lists for hashtable Jerry Stuckle <jstucklex@attglobal.net> - 2016-04-29 10:35 -0400
                          Re: Testing nodes and lists for hashtable Ken Brody <kenbrody@spamcop.net> - 2016-04-29 11:29 -0400
                            Re: Testing nodes and lists for hashtable Jerry Stuckle <jstucklex@attglobal.net> - 2016-04-29 12:21 -0400
                              Re: Testing nodes and lists for hashtable Ken Brody <kenbrody@spamcop.net> - 2016-05-02 11:55 -0400
                                Re: Testing nodes and lists for hashtable Jerry Stuckle <jstucklex@attglobal.net> - 2016-05-02 13:13 -0400
                  Re: Testing nodes and lists for hashtable supercat@casperkitty.com - 2016-03-30 08:15 -0700
                    Re: Testing nodes and lists for hashtable Jerry Stuckle <jstucklex@attglobal.net> - 2016-03-30 11:39 -0400
                      Re: Testing nodes and lists for hashtable supercat@casperkitty.com - 2016-03-30 09:47 -0700
                  Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 08:48 -0700
            Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-30 02:18 -0700
              Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 06:56 -0700
                Re: Testing nodes and lists for hashtable Keith Thompson <kst-u@mib.org> - 2016-03-30 09:04 -0700
                  Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 09:25 -0700
                Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-04-01 00:52 -0700
        Re: Testing nodes and lists for hashtable Alla _ <modelling.data@gmail.com> - 2016-03-30 02:05 -0700
          Re: Testing nodes and lists for hashtable Geoff <geoff@invalid.invalid> - 2016-03-30 07:00 -0700

Page 12 of 17 — ← Prev page 1 … 10 11 [12] 13 14 … 17  Next page →


#84297

FromBen Bacarisse <ben.usenet@bsb.me.uk>
Date2016-03-18 01:05 +0000
Message-ID<87r3f8vbmp.fsf@bsb.me.uk>
In reply to#84293
Jerry Stuckle <jstucklex@attglobal.net> writes:

> On 3/17/2016 7:46 PM, Ben Bacarisse wrote:
<snip>
>> I'm not persuaded by any of this, but I'm not sure there's much value in
>> going further.  High load factors are never a good idea.

> High load factors are seldom a good idea, but almost impossible to avoid
> without wasting money on huge hardware systems.

We may have crossed wires.  I mean hash-table load factors which measure
in some way (I think there's more than one definition) the number of
hash collisions.  When using chaining (i.e. lists) rather than probing
you can almost always avoid high load factors because a bigger table
will use no more memory than a loaded one with chaining (though there
are fragmentation issues of course).

-- 
Ben.

[toc] | [prev] | [next] | [standalone]


#84300

FromJerry Stuckle <jstucklex@attglobal.net>
Date2016-03-17 22:35 -0400
Message-ID<ncfpc6$lbj$1@jstuckle.eternal-september.org>
In reply to#84297
On 3/17/2016 9:05 PM, Ben Bacarisse wrote:
> Jerry Stuckle <jstucklex@attglobal.net> writes:
> 
>> On 3/17/2016 7:46 PM, Ben Bacarisse wrote:
> <snip>
>>> I'm not persuaded by any of this, but I'm not sure there's much value in
>>> going further.  High load factors are never a good idea.
> 
>> High load factors are seldom a good idea, but almost impossible to avoid
>> without wasting money on huge hardware systems.
> 
> We may have crossed wires.  I mean hash-table load factors which measure
> in some way (I think there's more than one definition) the number of
> hash collisions.  When using chaining (i.e. lists) rather than probing
> you can almost always avoid high load factors because a bigger table
> will use no more memory than a loaded one with chaining (though there
> are fragmentation issues of course).
> 

Ah, OK, yes, we did cross wires.  But a good hash algorithm will
minimize collisions; it pays to have a good one.  OTOH, a binary tree
(or better, an AVL tree) will be faster to search than any but the
smallest lists, and are not that hard to implement.

-- 
==================
Remove the "x" from my email address
Jerry Stuckle
jstucklex@attglobal.net
==================

[toc] | [prev] | [next] | [standalone]


#84295

FromMalcolm McLean <malcolm.mclean5@btinternet.com>
Date2016-03-17 17:19 -0700
Message-ID<04152d6c-d0b5-454c-bba8-ebe4dbd8d4d8@googlegroups.com>
In reply to#84287
On Thursday, March 17, 2016 at 11:46:37 PM UTC, Ben Bacarisse wrote:
> Malcolm McLean <malcolm.mclean5@btinternet.com> writes:
> 
> > On Thursday, March 17, 2016 at 3:00:01 AM UTC, Ben Bacarisse wrote:
> >> Malcolm McLean <malcolm.mclean5@btinternet.com> writes:
> >> 
> >> I don't follow this.  You seem to be do some sort of worst-case
> >> analysis and I'm not sure why.
> >> 
> >> Given the hand-waving you are prepared to accept, I'm not sure there's
> >> any reason to think the distribution is significant.
> >> 
> > The question is the load factor at which a balanced binary search tree starts
> > to pay off. We know that a binary search tree will be better if the list of
> > items in the bucket can get very long.
> > So the first question is "how long are the lists likely to get?" - the answer is
> > approximated by the Poisson distribution (except in cases where the number
> > of buckets is so low that contents of one bucket give us reasonable information
> > about the likely contents of another).
> 
> It's approximated by many distributions.  In most case the approximation
> will be at it's worst in the tails which, for some reason I don't
> follow, you are focusing on.
> 
We know that a balanced binary search tree will have maximum advantage
over a linked list when the lists are long. The question is "at what 
load factor doe s a balanced binary tree become advantageous?". So
a reasonable way to answer that is to say "we must build the church 
Easter Sunday" - performance in the worst case is what is important to
us. But "worst case" has to have some statistical limit - obviously it 
is possible for a million items to hash to the same bucket out of
a million buckets, but, assuming a uniform hash function, it is far
more likely that the computer will develop an electrical fault. So
the question is, given a million items, a million buckets, and a 
million runs, what is the fullest bucket we are likely to see?

Then take the worst case and see what win a balanced binary tree
gives you over a linear search. 
>
> > So let's take a worst case scenario -
> > given a load factor, what's the longest bucket we're likely to see?
> 
> But why?  Even the crudest guess at the distribution will tell us that
> the cases near the mean dominate and the tails contains very rare
> cases.  Why an almost worst case analysis?
> 
> By picking (as you did) a "once in a million case" you are doing
> something very odd.  It is neither a worst-case analysis, nor one based
> on the distribution of cases (i.e. an "average case" analysis).  You use
> the distribution to pick one case that is "quite bad" but, because of
> the rather poor approximation at this low level, one you don't really
> know how common it is.
> 
You're right. But it's a reasonable procedure. The worst case is a
million items in one bucket, but, assuming a uniform hash, we'll
never see that. On the other hand, we can easily have a thousand 
buckets which we replenish a thousand times in the course of the
program, so if p is 1.0e-6, we'll see that on almost every run.
Take pe-12, and you've got a reasonable value for "impossible".
But the analysis isn't very sensitive to the value you choose for
"impossible", as long as you don't go for extremes like p = 5%
or p = 1.0e-1000, it's only two or three items in the worst case 
bucket.  
> 
> I'm not persuaded by any of this, but I'm not sure there's much value in
> going further.  High load factors are never a good idea.
> 
You posed the question.
But I agree. In practical terms, keep the load factor below about 3.0,
ad a linear list will be fine.
 

[toc] | [prev] | [next] | [standalone]


#84299

FromBen Bacarisse <ben.usenet@bsb.me.uk>
Date2016-03-18 01:39 +0000
Message-ID<87lh5gva1m.fsf@bsb.me.uk>
In reply to#84295
Malcolm McLean <malcolm.mclean5@btinternet.com> writes:

> On Thursday, March 17, 2016 at 11:46:37 PM UTC, Ben Bacarisse wrote:
<snip>
>> I'm not persuaded by any of this, but I'm not sure there's much value in
>> going further.  [...]
>> 
> You posed the question.

Yes, and you have answered it to your satisfaction.  I don't think it
can be answered like this with so many assumptions and numbers picked
out of the air.

<snip>
-- 
Ben.

[toc] | [prev] | [next] | [standalone]


#84311

FromMalcolm McLean <malcolm.mclean5@btinternet.com>
Date2016-03-18 02:54 -0700
Message-ID<95250e05-1aaf-471a-aa98-771e7f921322@googlegroups.com>
In reply to#84299
On Friday, March 18, 2016 at 1:40:01 AM UTC, Ben Bacarisse wrote:
> Malcolm McLean <malcolm.mclean5@btinternet.com> writes:
> 
> > On Thursday, March 17, 2016 at 11:46:37 PM UTC, Ben Bacarisse wrote:
> <snip>
> >> I'm not persuaded by any of this, but I'm not sure there's much value in
> >> going further.  [...]
> >> 
> > You posed the question.
> 
> Yes, and you have answered it to your satisfaction.  I don't think it
> can be answered like this with so many assumptions and numbers picked
> out of the air.
> 
And the result is that a hash table isn't going to see any worthwhile advantage 
from anything more sophisticated than a linked list bucket for load factors
or up to 3, whilst you need to start looking seriously at optimising the lists
when load factors got up to about 10 or higher.
That's more or less in line with what I would have said as a guesstimate.

But some of parameters are unknown, you just have to pick reasonable
figures. But change them within reasonable limits, and you won't get
a very different result.

Is it worth knowing? I don't know.

[toc] | [prev] | [next] | [standalone]


#84319

FromStephen Sprunk <stephen@sprunk.org>
Date2016-03-18 09:08 -0500
Message-ID<nch1v3$ccj$1@dont-email.me>
In reply to#84311
On 18-Mar-16 04:54, Malcolm McLean wrote:
> On Friday, March 18, 2016 at 1:40:01 AM UTC, Ben Bacarisse wrote:
>> Malcolm McLean <malcolm.mclean5@btinternet.com> writes:
>>> Ben Bacarisse wrote:
>>>> I'm not persuaded by any of this, but I'm not sure there's much
>>>> value in going further.  [...]
>>>> 
>>> You posed the question.
>> 
>> Yes, and you have answered it to your satisfaction.  I don't think
>> it can be answered like this with so many assumptions and numbers
>> picked out of the air.
>> 
> And the result is that a hash table isn't going to see any worthwhile
> advantage from anything more sophisticated than a linked list bucket
> for load factors or up to 3, whilst you need to start looking
> seriously at optimising the lists when load factors got up to about
> 10 or higher. That's more or less in line with what I would have said
> as a guesstimate.

This means that if your load factor exceeds 3, you need to increase the
number of buckets to bring it below 3 again.  It doesn't mean that your
buckets need to be more efficient.

S

-- 
Stephen Sprunk         "God does not play dice."  --Albert Einstein
CCIE #3723         "God is an inveterate gambler, and He throws the
K5SSS        dice at every possible opportunity." --Stephen Hawking

[toc] | [prev] | [next] | [standalone]


#84338

FromMalcolm McLean <malcolm.mclean5@btinternet.com>
Date2016-03-18 11:20 -0700
Message-ID<04eb90ef-0799-40bb-b683-cbf945554c6e@googlegroups.com>
In reply to#84319
On Friday, March 18, 2016 at 2:08:46 PM UTC, Stephen Sprunk wrote:
> On 18-Mar-16 04:54, Malcolm McLean wrote:
> 
> > And the result is that a hash table isn't going to see any worthwhile
> > advantage from anything more sophisticated than a linked list bucket
> > for load factors or up to 3, whilst you need to start looking
> > seriously at optimising the lists when load factors got up to about
> > 10 or higher. That's more or less in line with what I would have said
> > as a guesstimate.
> 
> This means that if your load factor exceeds 3, you need to increase the
> number of buckets to bring it below 3 again.  It doesn't mean that your
> buckets need to be more efficient.
> 
It's not always possible to re-hash. Usually that's what you do, I agree.

[toc] | [prev] | [next] | [standalone]


#84345

FromBen Bacarisse <ben.usenet@bsb.me.uk>
Date2016-03-18 21:29 +0000
Message-ID<87bn6bv5j6.fsf@bsb.me.uk>
In reply to#84311
Malcolm McLean <malcolm.mclean5@btinternet.com> writes:

> On Friday, March 18, 2016 at 1:40:01 AM UTC, Ben Bacarisse wrote:
>> Malcolm McLean <malcolm.mclean5@btinternet.com> writes:
>> 
>> > On Thursday, March 17, 2016 at 11:46:37 PM UTC, Ben Bacarisse wrote:
>> <snip>
>> >> I'm not persuaded by any of this, but I'm not sure there's much value in
>> >> going further.  [...]
>> >> 
>> > You posed the question.
>> 
>> Yes, and you have answered it to your satisfaction.  I don't think it
>> can be answered like this with so many assumptions and numbers picked
>> out of the air.
>> 
> And the result is that a hash table isn't going to see any worthwhile advantage 
> from anything more sophisticated than a linked list bucket for load factors
> or up to 3, whilst you need to start looking seriously at optimising the lists
> when load factors got up to about 10 or higher.
> That's more or less in line with what I would have said as a
> guesstimate.

Yes, but I think you still have a guesstimate.  You picked numbered
based on what you think of a reasonable costs for linked list and binary
tree operations.  These unsurprisingly give an answer congruent with
your gut feelings.

> But some of parameters are unknown, you just have to pick reasonable
> figures. But change them within reasonable limits, and you won't get
> a very different result.

Yes, because reasonable guesses must be those that produce the sort of
answer you'd guess in the first place.

> Is it worth knowing? I don't know.

I don't know if it can be called knowing.

-- 
Ben.

[toc] | [prev] | [next] | [standalone]


#84346

FromMalcolm McLean <malcolm.mclean5@btinternet.com>
Date2016-03-18 17:48 -0700
Message-ID<0f687dc1-9b0c-438c-8967-d1036777a0de@googlegroups.com>
In reply to#84345
On Friday, March 18, 2016 at 9:29:42 PM UTC, Ben Bacarisse wrote:
> Malcolm McLean <malcolm.mclean5@btinternet.com> writes:
>
> > But some of parameters are unknown, you just have to pick reasonable
> > figures. But change them within reasonable limits, and you won't get
> > a very different result.
> 
> Yes, because reasonable guesses must be those that produce the sort of
> answer you'd guess in the first place.
>
There's an element of that.
But at least we state what the assumptions are. If you don't like 
them, plug in another value.
Of course if the values are way off base - e.g. binary search costs
a billion times more than linear search, you'll get radically 
different results. But few people would defend that sort of figure.
Exactly how far you penalise binary search is a bit of an ill-defined
question.
> 
> > Is it worth knowing? I don't know.
> 
> I don't know if it can be called knowing.
> 

[toc] | [prev] | [next] | [standalone]


#84210

FromStephen Sprunk <stephen@sprunk.org>
Date2016-03-16 22:09 -0500
Message-ID<ncd6uk$ipu$1@dont-email.me>
In reply to#84195
On 16-Mar-16 17:39, Malcolm McLean wrote:
> A balanced search tree finds the item in logarithmic time
> plus a constant, a linear one in linear time. So call
> the constant "c" and the scale "s", so time_bbst =
> log2(time_link) * s + c;
> 
> We'll assume s is 2 and c is 1 (a bit of fudge, to allow for
> time to balance the bbst)
> 
> Now the tree is twice as fast as the linked list at
> N = 20 or so, and four times as fast at N = 40.
> 
> So the answer, with admittedly a lot of hand-waving, is that
> the balanced binary tree starts becoming worth it for a load 
> factor of about 3.0 or so, and will increasingly start to
> be the method of choice for a load factor of 10 or so.

*scratches head*

Isn't a heavily loaded bucket balanced out by a lightly loaded bucket,
i.e. the same result as if the buckets were all of average load?

Given all the work (C) to keep a tree balanced, you'd need a large N for
O(log N)+C to beat O(N) in practice; you can cut the load factor by
increasing the number of buckets, so N should never _be_ large.

S

-- 
Stephen Sprunk         "God does not play dice."  --Albert Einstein
CCIE #3723         "God is an inveterate gambler, and He throws the
K5SSS        dice at every possible opportunity." --Stephen Hawking

[toc] | [prev] | [next] | [standalone]


#84237

Fromraltbos@xs4all.nl (Richard Bos)
Date2016-03-17 11:49 +0000
Message-ID<56ea995a.1483062@news.xs4all.nl>
In reply to#84210
Stephen Sprunk <stephen@sprunk.org> wrote:

> On 16-Mar-16 17:39, Malcolm McLean wrote:

> > So the answer, with admittedly a lot of hand-waving, is that
> > the balanced binary tree starts becoming worth it for a load 
> > factor of about 3.0 or so, and will increasingly start to
> > be the method of choice for a load factor of 10 or so.
> 
> *scratches head*
> 
> Isn't a heavily loaded bucket balanced out by a lightly loaded bucket,
> i.e. the same result as if the buckets were all of average load?

Not really; to begin with, a heavily loaded bucket is (ipso facto) hit
much more often than a lightly loaded one, and therefore has a greater
effect on efficiency.

> Given all the work (C) to keep a tree balanced, you'd need a large N for
> O(log N)+C to beat O(N) in practice; you can cut the load factor by
> increasing the number of buckets, so N should never _be_ large.

Which is why _almost_-balanced trees are more often used, IME, than
perfectly balanced ones.

Richard

[toc] | [prev] | [next] | [standalone]


#84252

FromMalcolm McLean <malcolm.mclean5@btinternet.com>
Date2016-03-17 08:45 -0700
Message-ID<f98e0681-77a1-42df-b949-559b2a384c02@googlegroups.com>
In reply to#84237
On Thursday, March 17, 2016 at 11:49:43 AM UTC, Richard Bos wrote:
> Stephen Sprunk <stephen@sprunk.org> wrote:
> 
> > On 16-Mar-16 17:39, Malcolm McLean wrote:
> 
> > > So the answer, with admittedly a lot of hand-waving, is that
> > > the balanced binary tree starts becoming worth it for a load 
> > > factor of about 3.0 or so, and will increasingly start to
> > > be the method of choice for a load factor of 10 or so.
> > 
> > *scratches head*
> > 
> > Isn't a heavily loaded bucket balanced out by a lightly loaded bucket,
> > i.e. the same result as if the buckets were all of average load?
> 
> Not really; to begin with, a heavily loaded bucket is (ipso facto) hit
> much more often than a lightly loaded one, and therefore has a greater
> effect on efficiency.
> 
Then generally you assume the worst case - that all the entries hash to
the same bucket for example. However assuming a good hash function and
non-malciious input, that's less likely than that the computer will break.
So you assume the worst halfway plausible case, which is that we have
something like one in a million bad luck, we've got a bad bucket and an
item we need to constantly look up is in that bucket.
So I chose 1.0e-12 (a thousand buckets, a thousand passes, a one in a
million cut-off for let's say it will never happen ). But it's hand-waving.
However the probabilities drop off very quickly, so it doesn't make too 
much difference if you say "a thousand passes" or "a million passes".
>
> > Given all the work (C) to keep a tree balanced, you'd need a large N for
> > O(log N)+C to beat O(N) in practice; you can cut the load factor by
> > increasing the number of buckets, so N should never _be_ large.
> 
> Which is why _almost_-balanced trees are more often used, IME, than
> perfectly balanced ones.
> 
One question is what penalty to apply to the balanced binary search tree
over the linked list. Another is what's the relative importance of expected
versus worst-case performance. Sometimes the worst case dominates
- do you build the church for Easter Sunday? 

[toc] | [prev] | [next] | [standalone]


#84264

FromStephen Sprunk <stephen@sprunk.org>
Date2016-03-17 12:11 -0500
Message-ID<nceoaq$p3u$1@dont-email.me>
In reply to#84237
On 17-Mar-16 06:49, Richard Bos wrote:
> Stephen Sprunk <stephen@sprunk.org> wrote:
>> On 16-Mar-16 17:39, Malcolm McLean wrote:
>>> So the answer, with admittedly a lot of hand-waving, is that the
>>> balanced binary tree starts becoming worth it for a load factor
>>> of about 3.0 or so, and will increasingly start to be the method
>>> of choice for a load factor of 10 or so.
>> 
>> Isn't a heavily loaded bucket balanced out by a lightly loaded
>> bucket, i.e. the same result as if the buckets were all of average
>> load?
> 
> Not really; to begin with, a heavily loaded bucket is (ipso facto)
> hit much more often than a lightly loaded one, and therefore has a
> greater effect on efficiency.

Doh!  That makes sense, thanks.

>> Given all the work (C) to keep a tree balanced, you'd need a large
>> N for O(log N)+C to beat O(N) in practice; you can cut the load
>> factor by increasing the number of buckets, so N should never _be_
>> large.
> 
> Which is why _almost_-balanced trees are more often used, IME, than 
> perfectly balanced ones.

Still, unless the input is truly disordered, there will be overhead in
having to balance the tree, which offsets the lower search time.  In
Alla's case, we know the input is ordered, so without that step, the
tree degenerates to a linked list.

S

-- 
Stephen Sprunk         "God does not play dice."  --Albert Einstein
CCIE #3723         "God is an inveterate gambler, and He throws the
K5SSS        dice at every possible opportunity." --Stephen Hawking

[toc] | [prev] | [next] | [standalone]


#84267

Fromsupercat@casperkitty.com
Date2016-03-17 10:31 -0700
Message-ID<d9713cf6-9e9b-4014-aa93-432805ba7cdd@googlegroups.com>
In reply to#84264
On Thursday, March 17, 2016 at 12:12:00 PM UTC-5, Stephen Sprunk wrote:
> Still, unless the input is truly disordered, there will be overhead in
> having to balance the tree, which offsets the lower search time.  In
> Alla's case, we know the input is ordered, so without that step, the
> tree degenerates to a linked list.

If one has a known number of items in a sorted list, a perfectly-balanced
tree can be readily constructed in linear time; this can be done especially
nicely if the source list supports random access, but random access isn't
required to make the technique work.

[toc] | [prev] | [next] | [standalone]


#84275

FromMalcolm McLean <malcolm.mclean5@btinternet.com>
Date2016-03-17 11:09 -0700
Message-ID<1306effe-c277-4b2b-baca-ac8f6edc2d7e@googlegroups.com>
In reply to#84267
On Thursday, March 17, 2016 at 5:32:05 PM UTC, supe...@casperkitty.com wrote:
> On Thursday, March 17, 2016 at 12:12:00 PM UTC-5, Stephen Sprunk wrote:
> > Still, unless the input is truly disordered, there will be overhead in
> > having to balance the tree, which offsets the lower search time.  In
> > Alla's case, we know the input is ordered, so without that step, the
> > tree degenerates to a linked list.
> 
> If one has a known number of items in a sorted list, a perfectly-balanced
> tree can be readily constructed in linear time; this can be done especially
> nicely if the source list supports random access, but random access isn't
> required to make the technique work.
>
If we know that the table will be set up with a list of sorted keys, then we
can do a binary search instead of a hash, and get O log N performance
with zero overhead, and that will generally be our method of choice.

[toc] | [prev] | [next] | [standalone]


#84284

FromStephen Sprunk <stephen@sprunk.org>
Date2016-03-17 16:15 -0500
Message-ID<ncf6j4$ls7$1@dont-email.me>
In reply to#84275
On 17-Mar-16 13:09, Malcolm McLean wrote:
> supe...@casperkitty.com wrote:
>> Stephen Sprunk wrote:
>>> Still, unless the input is truly disordered, there will be 
>>> overhead in having to balance the tree, which offsets the lower 
>>> search time.  In Alla's case, we know the input is ordered, so 
>>> without that step, the tree degenerates to a linked list.
>> 
>> If one has a known number of items in a sorted list, a 
>> perfectly-balanced tree can be readily constructed in linear time; 
>> this can be done especially nicely if the source list supports 
>> random access, but random access isn't required to make the 
>> technique work.
> 
> If we know that the table will be set up with a list of sorted keys, 
> then we can do a binary search instead of a hash, and get O log N 
> performance with zero overhead, and that will generally be our
> method of choice.

With a reasonable number of buckets, i.e. low load factor, searching a
hash table should be O(1), which beats O(log N).  That's the point!

S

-- 
Stephen Sprunk         "God does not play dice."  --Albert Einstein
CCIE #3723         "God is an inveterate gambler, and He throws the
K5SSS        dice at every possible opportunity." --Stephen Hawking

[toc] | [prev] | [next] | [standalone]


#84296

FromBen Bacarisse <ben.usenet@bsb.me.uk>
Date2016-03-18 00:39 +0000
Message-ID<87wpp0vcur.fsf@bsb.me.uk>
In reply to#84275
Malcolm McLean <malcolm.mclean5@btinternet.com> writes:

> On Thursday, March 17, 2016 at 5:32:05 PM UTC, supe...@casperkitty.com wrote:
>> On Thursday, March 17, 2016 at 12:12:00 PM UTC-5, Stephen Sprunk wrote:
>> > Still, unless the input is truly disordered, there will be overhead in
>> > having to balance the tree, which offsets the lower search time.  In
>> > Alla's case, we know the input is ordered, so without that step, the
>> > tree degenerates to a linked list.
>> 
>> If one has a known number of items in a sorted list, a perfectly-balanced
>> tree can be readily constructed in linear time; this can be done especially
>> nicely if the source list supports random access, but random access isn't
>> required to make the technique work.
>>
> If we know that the table will be set up with a list of sorted keys, then we
> can do a binary search instead of a hash, and get O log N performance
> with zero overhead, and that will generally be our method of choice.

Given that picture why not use a minimal perfect hash function?  The
cost is compile-time and execution is O(1).

-- 
Ben.

[toc] | [prev] | [next] | [standalone]


#84141

FromStephen Sprunk <stephen@sprunk.org>
Date2016-03-16 12:04 -0500
Message-ID<ncc3hj$iqd$1@dont-email.me>
In reply to#84100
On 16-Mar-16 09:05, Malcolm McLean wrote:
> On Wednesday, March 16, 2016 at 1:46:27 PM UTC, Ben Bacarisse wrote:
>> I've not seen the acronym BBST before but I presume you mean
>> balanced binary search tree?  I'd might try a binary tree, but I am
>> not sure it's worth balancing it. 

When reading in a sorted dictionary, an unbalanced tree devolves into a
linked list.  However, when loading data in bulk, it's probably better
to have a single balancing pass at the end rather than a tree that
self-balances after each node is added.

>> I wonder at what load factor the balanced tree starts to pay off.
> 
> Assuming a perfectly random hash-funtion, [math stuff]
> 
> That tells you the highest number of items in a bucket you can
> reasonably expect. You will then hit that bucket Nworst/Nbuckets
> times.
> 
> So that tells you if a more sophisticated method than a linked list
> is going to be worth it.

OTOH, if you have so many nodes per bucket that the data structure
within buckets actually matters, you're almost certainly better off
increasing the number of buckets until it doesn't anymore.

S

-- 
Stephen Sprunk         "God does not play dice."  --Albert Einstein
CCIE #3723         "God is an inveterate gambler, and He throws the
K5SSS        dice at every possible opportunity." --Stephen Hawking

[toc] | [prev] | [next] | [standalone]


#84259

FromSteve Thompson <stevet810@gmail.com>
Date2016-03-17 14:21 +0000
Message-ID<lSf9Or.0eP.ZJtGh@gmail.com>
In reply to#84097
On Wed, Mar 16, 2016 at 01:46:16PM +0000, Ben Bacarisse wrote:
> Richard Heathfield <rjh@cpax.org.uk> writes:
> > 1) a module for storing text data, and for destroying it when no
> > longer needed, with the 'constructor' and 'destructor' adhering to a
> > generic, flexible format;
> > 2) a module for storing /arbitrary/ data in a linked list, for
> > searching that list, for destroying specific nodes in that list, and
> > for destroying the whole list;
> > 3) a module for calculating a hash from arbitrary data;
> > 4) a module for building, using, and destroying a hash table with
> > arbitrarily many buckets.
> > 5) main(), which would have been trivially simple to write by this stage.
> >
> > Except that I'd have used a BBST in each bucket, rather than a linked
> > list.
> 
> I've not seen the acronym BBST before but I presume you mean balanced
> binary search tree?  I'd might try a binary tree, but I am not sure it's
> worth balancing it.  At worse you'd get a list, and since the plan is
> usually to keep the lists short, would it pay off?  Have done any tests
> of hashing + list, hashing + tree and hashing + some balanced tree?  I
> wonder at what load factor the balanced tree starts to pay off.

And you call yourself a programmer?



Regards,

Steve Thompson

[toc] | [prev] | [next] | [standalone]


#84073

FromAlla _ <modelling.data@gmail.com>
Date2016-03-16 00:53 -0700
Message-ID<a4951e79-1687-455f-9350-6dd143a1d89d@googlegroups.com>
In reply to#84044
<snip>
> > bool unload(void)
> > {
> >   bool had_to_unload = false;
> >
> >   if(hashtable != NULL)
> >   {
> >     unload_nodes();
> >     free(hashtable);
> >     hashtable = NULL; /* ensure that a second call to unload()
> >                          will stop at the test. */
> >     had_to_unload = true;
> >   }
> >   return had_to_unload;
> > }
> 
> I don't see what the bool return is for, but I may well have missed some
> posts as the program developed.

It's a prerequisite condition - I can't change function declarations;
and I assume the staff uses this to test if a student has correctly 
implemented unload() function. Not sure, though. 
> 
><snip>

[toc] | [prev] | [next] | [standalone]


Page 12 of 17 — ← Prev page 1 … 10 11 [12] 13 14 … 17  Next page →

Back to top | Article view | comp.lang.c


csiph-web