Re: Look at these update from M$ Corporation.
On Sat, Aug 02, 2003 at 04:10:00AM -0700, Steve Lamb wrote:
> On Sat, 2 Aug 2003 11:14:27 +0200
> David Fokkema <dfokkema@ileos.nl> wrote:
> > On Fri, Aug 01, 2003 at 04:11:28PM -0500, Manoj Srivastava wrote:
> > > Because we chose not to? Because we have objections in
> > > principle to people offloading their spam fighting effort to innocent
> > > correspondents? Because I have better things to do than to beg and
> > > plead for my answers to be heard by people who asked for help in the
> > > first place?
>
> > Then send your answer to the list.
>
> Sometimes correspondence is not appropriate to send through the list but
> is still relevant. For example when Alan Conner objected to PGP signatures I
> was looking through the signed messages and noticed one from Manoj had a bad
> signature. I double-checked to ensure the keys I had were the correct ones
> and even went so far as to retrieve them again from a keyserver. Still a bad
> signature. I sent a copy of the message to Manoj with an explanation of the
> behavior in case something was misconfigured on his(?) end. Certainly not
> something I want to bother the list with and quite appropriate to send
> directly. Not worth enough effort to hit reply on a C-R, however.
Got your point. However, it is still a matter of principle. I suspect
that if people didn't _really_ mind C-R, they would consider simply
replying to be a lot less effort than retrieving the keys, checking
them, writing the mail and alike. Then it would only be a marginally
small amount of extra effort. Apparently, many people really dislike C-R
on principle. Well, I don't want to convert people, I was just more or
less curious about the reasons some people refuse to reply to a C-R.
>
> > Getting tens of mails of spam a day and hitting `d' on them requires a
> > lot more brain cycles then (almost) blindly responding to a challenge.
> > But that is just my guess, or my opinion.
>
> Then install SA and spend a few minutes perusing the configuration
> options. I find SA quite effective with the Bayesian filters turned on and
> set to the default auto-learn limits. I also add in a Vipul's Razor check.
> About 15 minutes of work and I cut my spam load from 10-20 a day to 1-2 a
> week. Here's the kicker. I'm not annoying dozens of other people on an
> ongoing basis to do it.
I have installed SA and I have spent a few minutes for the
configuration. I have only seen a few false positives from M$ on my
hotmail account which I, after reading them through and checking the SA
report, considered to be spam after all. I have only had a few false
negatives. All in the course of a few months. All in all, I think that
filtering techniques are very useful.
The problem is, is that I generally like the idea of C-R and
whitelisting. In my opinion, it is simple and elegant. It doesn't
require lots of CPU cycles or man hours maintaining filters and it
doesn't have false positives or negatives (in theory, anyway). The only
requirement (and drawback) is that other people reply to a C-R from time
to time. If configured friendly, only one time for each new person you
start mailing. I myself don't think this is annoying. The only reason I
have not installed tmda and use it myself, is the feeling I had that
_some_ people actually _might_ be annoyed. This feeling turned out to be
right, at least on this list. But I don't really understand the hard
feelings and strong emotional currents involved.
David
Reply to: