[cryptography] Nonlinear bias in subscription state. Re: not unsubscribing (Re: Unsubscribing)

Marsh Ray marsh at extendedsubset.com
Thu Jun 16 17:15:00 EDT 2011

On 06/16/2011 02:17 PM, Adam Back wrote:
> Trust me the noise level on here is zero compared to usenet news
> flame fests, spam, DoS etc. The maintainer is removing spam for one
> (I think).

Anything looks acceptable if you're willing to set your standard of
comparison low enough.

Many of us aren't on those lists for exactly that reason. I miss usenet. :-(

> Personally I find it kind of annoying when people want to squelch any
> interesting discussion about societal implications as that is part of
> what is interesting to me about crypto.

This is a really fair point and seems related to what Ian G said a few
days ago:
> Crypto people spend all their lives learning theoretical crypto in
> groups like this.  Then they go and apply their theoretical crypto
> out in the real world, and it bombs.  Or worse: [...]

We all have our own ideal list. Often what we want is support for our
own ideas and more about the same stuff we already like. It's probably
better when we don't get exactly what we want in this respect. But it
also shouldn't be so different that valued contributors wander off.

On 06/16/2011 02:17 PM, Adam Back wrote:
> Those discussions often are directly or indirectly actually threat
> model and design consideration discussions for privacy technology
> protocols, and about the long term deployment prospects of different
>  designs.

Great! Doesn't that imply there's a design principle that's transferable
to other contexts where crypto is used?

If so, couldn't we do a better job of abstracting out the lessons for
cryptosystems in general?

For example, data security and crypto attacks are often discussed in
terms of "value of asset being protected" relative to "costs imposed
upon the attacker". These values may even assume units of an actual
currency. We ought to be able to discuss such things without it ending
up down the rathole of government monetary policy.

> I never liked Perry's list much because before much could said
> someone would complain that there was no crypto-math in the post. I
> can do crypto-math as well as the next guy, but thats not only why
> I'm on this list.

Sometimes that kind of thing is useful as a bozo filter, one which I
would regularly fail. But $deity save us all from those bozos that are
armed with crypto-math.

> Its fairly easy to bulk skip a thread by subject line also. Press
> 'N' as they used to say in the days of tin or whatever it was :)

One can also set a filter to send certain author's posts to the trashcan
too. It can make a list such as this one much nicer. ;-)

- Marsh

More information about the cryptography mailing list