Gnupg good for big groups?

Robert J. Hansen rjh at sixdemonbag.org
Sun Aug 8 18:44:21 CEST 2010


On 8/8/2010 10:49 AM, MFPA wrote:
> How many of these 22 were within the first week or so?
> I find very few messages not encrypted to mine.

Again, network theory to the rescue.  Generally speaking, nodes that
carry little traffic are responsible for more problems than those that
carry a lot.  There are of course exceptions.

> It is. I'm quite surprised at the proportion of unencrypted messages,
> and at the proportion of members not encrypting to somebody's key. I
> would hope that latter figure dropped significantly if non-encryption
> to keys posted within the last week were disregarded.

Right, but at that point you're coming close to cherrypicking --
disregarding data points in order to reach a result that's more in line
with your preconceptions.  Nobody ever wakes up and says, "today I think
I'll cherrypick."  It's almost always a subconscious process: "well, I
can disregard that data, it's clearly anomalous because..."

Sometimes, data really /is/ anomalous and has to be thrown out.  If you
were to pick a random sample of 100 people and come to an average of
their incomes, you should disregard Bill Gates if he happens to be in
your data set.  More often than not, though, anomalous data isn't
anomalous at all -- it just illuminates a part of the problem you didn't
know existed.




More information about the Gnupg-users mailing list