team-karma team mailing list archive

["Prasad H. L." <hlprasu@xxxxxxxxx>]

Of course with sample size higher the uniformity is projected better.
But, the point is that we should not use such a technique when the
sample size is small (~2-3).

2009/9/18 arvind iyer <iyer.arvind.sundaram@xxxxxxxxx>:
> Ran the following
>
> for i in `seq 1 1600`;do cat /proc/sys/kernel/random/uuid |md5sum ;done|gawk
> '{count[substr($1,1,1)]++}END{for (i in count) {print
> i,"=>",count[i]}}'|sort
>
> to give me a distribution of the first character in md5
>
> i got the following samples
>
> for 160,
> 0 => 11
> 1 => 14
> 2 => 9
> 3 => 5
> 4 => 6
> 5 => 6
> 6 => 15
> 7 => 10
> 8 => 10
> 9 => 9
> a => 13
> b => 13
> c => 11
> d => 9
> e => 11
> f => 8
>
> for 1600
> 0 => 97
> 1 => 97
> 2 => 96
> 3 => 100
> 4 => 93
> 5 => 112
> 6 => 105
> 7 => 107
> 8 => 91
> 9 => 105
> a => 112
> b => 91
> c => 96
> d => 101
> e => 109
> f => 88
>
>
> and for 16000,
>
> 0 => 959
> 1 => 1039
> 2 => 1062
> 3 => 969
> 4 => 1019
> 5 => 1044
> 6 => 1042
> 7 => 972
> 8 => 960
> 9 => 985
> a => 978
> b => 1006
> c => 1026
> d => 975
> e => 984
> f => 980
>
>
> etc
>
> You may try it, and see if they are mixed enough. I guess not bad at all!
> The effectiveness grows with the size.
>
>
>
> For ppl other than prasad, i know that the context is missing. please look
> out for a mail from me soon regarding ASDFS
>
>
> ----------------------------------------------------------
>
>
>
>
>
>
>
>
> --
> ================================
> Work while you are alive, you can rest later
>
>
>
> --
> ================================
> Work while you are alive, you can rest later
>
> _______________________________________________
> Mailing list: https://launchpad.net/~team-karma
> Post to     : team-karma@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~team-karma
> More help   : https://help.launchpad.net/ListHelp
>
>