Selection differences -> adds and removes

[Lusercop]

On Thu, Nov 17, 2005 at 11:51:51AM +0000, Jonathan Peterson wrote:
>> Sorry about the wide copy here, but we had this one LESS THAN A WEEK AGO!
>> Has the level of technical competence really declined this far!?
> If someone wants to un-sub, they are no longer interested in the list 
> contents. So they have stopped reading it, so they have no idea what was 
> posted five days, or indeed five minutes ago.

How helpful of them. So, knowing that they're doing the wrong thing "sorry
for the wide copy", they go ahead and do it anyway. Note that I didn't
mention the other two obvious annoyances, top-posting, and replying to a
message with completely different relevance.

> Eventually, for some reason, they stop deleting the irrelevant emails and 
> instead decide to un-subscribe. So they look for the unsub info in the 
> message footer and don't find it. Then they ask to be unsubscribed.

Fair enough - ish.

> A fail to see anything but blinding common sense in this approach. I 
> realise that the headers contain unsubscribe information. If there was 
> some key information that I really wanted to make available to list users, 
> the message *headers* wouldn't actually be my first port of call, but hey, 
> that decision was doubtless made aeons ago by l33t *nix gods who had never 
> had trouble unsubscribing from a list in their lives, and so had no idea 
> how best to solve the problem.

The "l33t *nix gods" of which you speak are none such. However, I would expect
someone who signs himself off as:
"Computing Technical Services (CTS) - Unix Admin EU"
to have at least a basic understanding of mail headers and some basic level
of technical competence. If you have a problem with this approach, perhaps
you should be participating in the relevant IETF WG. (it'll be some list
@imc.org, I'm sure).

> "These Lusers can't unsub"

No, "these people are lusers because they know they're doing something wrong
and do it anyway - posting to the whole list".

> "Yeah, man, it's like - it's like it's always September on the internet 
> now"

It's been September for well over a decade.

[snip the rest of the diatribe which is a lovely beautiful parody, but
 basically nothing like how the real mindset which causes people like me
 to be irritated about such things actually works, and is done in l33t5p34k]

> Meanwhile, while _your_ email may no longer be any use for business 
> purposes, mine certainly is. I use normal Bayesian anti-spam, auto-delete 
> the auto-generated name spams, and get about 2-5 spams a day coming 
> through. I can conduct business on that basis.

Fair enough. I can conduct most business I need to conduct. My email is
useable, and became so once I took technical measures to cut down the
spam. My point stands. Unfiltered email - which is what you need to
guarantee that you'll never lose an email ever - is no longer useable.

My measure of whether a site is competent or not works for me. One
interesting one that I got fed up of replying to is SingNet in Singapore
who keep getting blacklisted, by their relays sending mail to addresses
advertised in web pages as "please do not send mail to this address" in
order to pick up automated harvesters, they kept sending me a mail saying
that I could "uplift the blocking", because they'd closed all ways and
means of relaying. The point is, I'm still getting spam from them, so they
got blocked. I know it's spam, because I've never used these addresses in
anything other than that context.

The rDNS checks, header syntax and other checks which I assume Jens was
referring to, I very much believe in, as incorrect syntax could cause
unwanted behaviour in my mail client, or indeed in the mail storage
system. I want to reject incorrect syntax (well-formedness of messages on
the internet is very important, IMO), on the basis that it's a security
risk. The rDNS thing is a bit different, but, if you have a reverse record
you shouldn't expect anything to work, unless the forward and reverse data
should be able to tally. Other things than mail will reject on that basis.
Merely having a reverse record at all is a good indicator that you are
concerned about accountability - and hence that you should be trusted.

When it comes to things like dialup lists and dialup/DSL like names, let's
be fair here. Much of the spam I see comes from zombie windows machines
with "dsl" or "dial" in their names. There is little reputation associated
with these, the systems can move around (in a dynamic pool) and other such.

If you run an open relay on today's internet, you also deserve everything
you get, and yes, I include John Gilmore in this.

If on top of all of that, you need the rest of SA and/or Bayesian filtering,
then I'd say your email has become unusable. Can you guarantee that you've
never lost an important mail due to your filtering?

> My domain is about 5-6 years old, and the problem is not getting worse, if
> anything it improves slightly over time. Oddly enough, my work email,
> which I am far more profligate with, has even less spam than my private
> one - although I've had that for 4-5 years as well.

That's interesting, and unusual. I'd say it takes 6 months to a year (from
some experimentation) for addresses which get published to get onto a spam
list, but I don't think any of them have really failed to get listed. I have
a huge spam corpus from replies to my list postings (which are filtered such
that I only read those addresses if they came from the list).

>> Grngh
> And finally, gentlemen, I will outrage the world, with my terrifying 5 
> LINE SIGNATURE OF DOOM. Or not.

I have seen a one line mail (that had no more actual content than the text
of the body) take up 200k on disc due to signatures, embedded images etc.
from our CEO.

Not long after that, we were asked to do some spring cleaning on the
company mailserver, as it was running out of disk.

While some of these old traditions are silly, I do still read my mail (and
particularly this list) in an 80x24 xterm, and so screenfuls of sig is
still irritating (although less so than when I was on dialup).

One of the clients at my old job used to send their emails with a 12 line
disclaimer as a header, and a 20-odd line footer with their contact info.

This actually made it very hard to find the content (especially in situations
where the content was only one line) in their mails, thereby decreasing the
usefulness of the medium.

You're entitled to your opinion, and I hope you feel that I'm entitled to
mine. I'm not going to stop filtering on technical violations of protocol
anytime soon. If you can't get the protocol right then why should I be
trusting you to deliver me something that I'd want to touch with anything
less than a 30ft barge pole?

-- 
Lusercop.net - LARTing Lusers everywhere since 2002