991215 Klaus Weide calculated:
> Let's say there are 20 temp files existing & kept track of by Lynx
> at any one time, realistic with source_cache, UI pages, downloads.
> What are the chances that, among 20 filenames drawn out of 10000
> (with repetition), >= 2 are the same? I get: approximately 190/10000.
> (Can anyone confirm, or is my math just way off? :) )
not "approximately", exactly (grin): 1 + 2 + ... + 19 .
> So, at any point in time, there's a 1.9 per cent chance
> that two files collide! Definitely not a negligible effect.
certainly not negligible. as a user, i prefer short temp-file names,
as i sometimes want to grab them -- eg Visited Links -- :
given that security is important only when Lynx is shared by users,
not for people who compile their own & use their own /tmp directory,
an option to choose security vs simplicity might be called for here.
also, aren't there other ways of getting a random seed,
eg using the output of some version of `who', which is unpredictable
due to its dependence on all the system's users' activities?
of course, any villain who knows continually
exactly what the user does exactly when s/he does it
is going to be able to outflank any security devices, even God's.
--
========================,,============================================
SUPPORT ___________//___, Philip Webb : [EMAIL PROTECTED]
ELECTRIC /] [] [] [] [] []| Centre for Urban & Community Studies
TRANSIT `-O----------O---' University of Toronto
