On Tue, 29 Jul 2014 14:52:30 +1000 "shmick@xxxxxxxxxx" <shmick@xxxxxxxxxx> allegedly wrote: > > Andrew Lewman wrote: > > > > So roughly, 9122:1624 or 18% if I did my top-of-the-head math > > right. And technically, this is just requests for the .asc file, no > > idea if the person/bot actually ran gpg --verify on it. > > > > nginx maybe ? > that's interestingly high - i would have put money on < 5% > of course, we can't measure/know the human verification (or can we) Is it safe to assume that a high proportion of those actually bothering to download the .asc will then go on to verify? If not, why would they bother to download? I have just checked on my tails mirror and I get the slightly depressing results below: cat tails.log.1 | grep tails-i386-1.1.iso | grep -v .sig | sort -t. +0 -3 -u | wc -l 1774 cat tails.log.1 | grep tails-i386-1.1.iso.sig | sort -t. +0 -3 -u | wc -l 12 which I make 0.68% Note that I have used a different approach to Andrew because I think his approach will overestimate the number of file retrievals. Certainly in my logs I see multiple parallel connections from the same IP address (with resulting HTTP status code 206 for partial content). My sort above should strip those out. If I use Andrew's approach and do not strip the multiple connections from the same address I get the even more depressing results below: cat tails.log.1 | grep tails-i386-1.1.iso | grep -v .sig | wc -l 8194 which would give only 0.15% bothering to check the signature. (Of course, since many connections to my mirror may come over Tor my sort will remove any which happen to pop out of the same exit - so the approach is far from perfect. That said, my awstats runs show answers consistent with the approach I have used.) Mick --------------------------------------------------------------------- Mick Morgan gpg fingerprint: FC23 3338 F664 5E66 876B 72C0 0A1F E60B 5BAD D312 http://baldric.net ---------------------------------------------------------------------
Attachment:
signature.asc
Description: PGP signature
-- tor-talk mailing list - tor-talk@xxxxxxxxxxxxxxxxxxxx To unsubscribe or change other settings go to https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-talk