[Samba] Solved for 1/1300?

Jeremy Allison jra at samba.org
Mon Oct 30 08:18:00 GMT 2006


On Sun, Oct 29, 2006 at 11:35:22AM -0500, dkemper at intellisys-is.com wrote:
> Hello, 
> 
> Discrete Mathematics Question; how did the samba developers conclude that the default mangling method "hash" results in a probability-of-name-clash of 1/1300? 
> 
> Please note, I am not looking for samba solutions here, just intensely curious on how this number was solved for. 
> 
> For instance, 
> 
> Q: What is considered the sample space of all possible file names here, were the "-" and "_" taken into account? 
> 
> Q: Are we considering all possible filename are equally likely to occur? 
> 
> Jeremy A, perhaps you know? 

Tridge did the original calculation, maybe he remembers.

Jeremy.


More information about the samba mailing list