Re: UDP/DNS in Emacs

[Per Abrahamsen]

Eli Zaretskii <address@hidden> writes:

>> From: Per Abrahamsen <address@hidden>
>> Date: Tue, 02 Apr 2002 14:34:51 +0200
>> 
>> Eli Zaretskii <address@hidden> writes:
>> 
>> > My point is that basing this on the sender's ISP is a bad idea, one
>> > that unduly discriminates people who don't have much choice but to use
>> > whatever ISPs are available to them.  
>> 
>> It is their problem.  Why should I suffer for it?
>
> Because it will become your problem when relevant mail is
> automatically junked based on the domain from which it comes.

Not junked, but moved to a lower priority folder.  

And yes, _that_ is my problem.  So that should be my choice how to
deal with it.  

>> If Gnus can improve my blacklist rules, I will have to spend less time
>> skimming the misc.misc folder, and junk less messages by accident.
>
> If I understand your habits, it also means that some non-junk mail
> will be read once a month instead of once a day.

Yes.  It is all a matter of how to optimize on the two variables:  

1) How much time do I have to spend manually sorting spam.
2) How many non-spam messages may accidentially be junked (or delayed).

I am the one to judge how to prioritize between these.  I find it
insulting that you want to deny me that ability, especially with the
justification that you can prioritize my time better than I can.

> I think your assumption about availablility of good ISPs needs some
> reality check.  How well are you familiar with the situation outside
> Western Europe and North America?

I assume it it is similar to question in Denmark in the past.  If we
were unhappy with our sole national ISP, our alternative was to use an
UUCP connection to a foreign ISP for mail.  

In any case: Automatically junking everything from outside Western
Europe and Northern America is the most popular alternative to the
blaklists.  It is extremely tempting for me to blacklist SE Asia,
where the fraction of relevant mail (outside my whitelists) to spam is
near zero.  For most users, that is true.

>> To get out of the database, all you had to close it.
>
> I know all about this; do you?  How many times, if at all, did you
> need or try to deal with these problems?

KVL has often been blacklisted, when someone inside have installed a
new "NT Server" with too many defauls on.

> The number of times I removed my ISP from the data base is greater
> than the number of words in this message.  It doesn't help: a few days
> after that, it's in the data base again.

So you had an open relay again.  Incompetent ISP.  I also know the
kind of things competent ISP's have to do to keep out of the open
releay blacklists.  Basically, they have to test whether their
customer is running an open relay before relaying their customers
mail.

As a side-effect, supporting open relay blacklists helps motivate
ISP's to act competently.

>> Open relayes was (and probably still is) the most efficient way for
>> spammers to propagate their messages, and blocking based on open
>> relays was (and probably still is) the most efficient way to block
>> spam.
>
> As usual, this is a question of striking the fine balance of
> stopping the guilty without unduly punishing the innocent.

Nope.  It is not a question of punishment.  It is a question of
finding the balance where e-mail regain most of its usefulness.

> It might be worth remembering that spam is defined based on its
> content, not on the server from which it comes.

Nope, that would be totally irrelevant to the question.