message, but also test

[Ryan Lackey]

I've always used a special vmware instance whenever I wanted to do
something anonymously, as I assumed my OS choices and customizations 
might give me away to the other end as well, especially for anything I
don't notice (a good attack would be trying a well-known 0-day OS
vulnerability on an interactive counterparty, seeing if his machine
drops off the net -- you could do this several times and figure out
which OS is running).  

Running a pretty much standard win2k/xp virtual machine with whatever
anonymity functionality running on the windows side OR the unix side 
seems to deal with that, as well as making it easier to sniff all
output from the machine in realtime to make sure nothing else is going
through.  And it would then be very obvious which nym you're using.

I don't think gmail would work well with a 500MB archive -- I've used
it, and their UI seems inefficient past a certain point.  

General web indexing of html archives works pretty well --
("from, "to" searching would be equivalent to author searches, and
keyword + or - terms can be done through a web ui as easily)

I think I'd also prefer being able to do LOCAL searches on most of the
topics on cypherpunks, vs. web searches on someone else's machine, so
I'll just let people download raw archives when reconstructed, with a
5Mbps cap or something.

I am not sure how safely spam can be filtered from the list -- each
person would have his own view of what is spam.  I think ideally there
would be all messages in the spool, broken up by month or quarter or
year, and then a way to run various filters on it, like kill certain
senders, kill keywords, "traditional" spamassassin/bayesian type
systems, etc.  Then the ability to search over any subset.

Google doesn't support enough negative keywords to do this, and it
would be computationally expensive to do the filtering and reindexing
in realtime for each user.

If you had all the spools yourself locally, you could run whatever
spam elimination you considered proper once, and then search index,
with negative keywords for specific senders (or just excise certain
senders permanently as spam)


Quoting An Metet <anmetet at freedom.gmsociety.org>:
> Ryan Lackey writes:
> 
> > I have two questions:
> >
> > 1) Does anyone have actual performance measurements of ZKS from when
> > it was operational/at peak, in terms of bandwidth, MTU, latency, and
> > jitter?  Is there a good way to quantify just how far from
> > "acceptable" it was?
> 
> I don't have any actual measurements, but as far as I can remember it
> worked pretty well, i.e., slower than usual but still pages would load
> within a few seconds.
> 
> Recently I've experimented with the onion routing system at
> www.freehaven.net/tor, and the main problem there is slow startup times
> setting up the first path.  Seems to take a couple of minutes sometimes
> for the first web page to download.  I think it's struggling to find a
> working path, or something.  But then after that the performance seems
> comparable to Freedom.
> 
> I'm a big user of anonymity systems, and the worst problem I've had with
> proxies is remembering who I am supposed to be at the time.  Several times
> with Freedom and more recently with other proxies, I have done stuff
> using my real name when I was in the mode where my nym was being used,
> and vice versa.  Oops.  That's a pretty big mistake to make and can
> totally destroy your pseudonymity, both at the time and throughout the
> past lifetime of the nym.
> 
> What I'd like would be some kind of big, glaring indication that I am
> in "anonymous" mode, like overlaying some kind of color display on the
> screen, or maybe a crawling animation around the edges, or something.
> I realize that this is out of scope for most efforts of this type, but
> from my experience it's a big problem.
> 
> > (I also subscribed the al-qaeda node, and will probably finish setting
> > up the spamfiltered version of the list, as well as passing the back
> > archives through the same archiving software as current archives, and
> > search-indexing them, next time I get bored)
> 
> Making your deep archive available in search-indexed form would be a
> great service, as would spam-cleaning your current one.  May you grow
> bored soon.
> 
> BTW, how big is the entire CP archive when compressed?  Would it fit into
> a gmail account that someone could set up and share the passphrase to?
> I'd pay quite a few bucks to have a copy of that on my disk.
> 
> _______________________________________________
> cypherpunks mailing list
> cypherpunks at cypherpunks.metacolo.com
> http://cypherpunks.metacolo.com/mailman/listinfo/cypherpunks

-- 
Ryan Lackey [RL960-RIPE AS24812]   ryan at venona.com   +1 202 258 9251
OpenPGP DH 4096: B8B8 3D95 F940 9760 C64B   DE90 07AD BE07 D2E0 301F
_______________________________________________
cypherpunks mailing list
cypherpunks at cypherpunks.metacolo.com
http://cypherpunks.metacolo.com/mailman/listinfo/cypherpunks