On 4/8/2012 9:03 AM, Darren New wrote:
>> That is what exceptions are for. You might still have to check the
>> trap, but
>> it would be a "slightly" smarter trap. Right now, the trap tries to
>> rely on
>> blacklist data, and keyword identification,
>
> If you want to do it at the ISP level, you can't really do a very good
> job of keyword matching. Maybe you really *do* buy your viagra from an
> online pharmacy. How do you check the trap if some other ISP has thrown
> away the email before it even gets to you?
>
Actually, it is done on the ISP level, but its shoved into the "spam" 
folder, when things work right. The problem is, it almost never works 
right. lol

>> "viagra", and if a few other words are there, its flagged, hence the
>> moronic
>> fact that those slip through, while Hotmail has **multiple** times
>> actually
>> flagged legit emails Origin, about things going on with Star Wars: KOTOR.
>
> And that's the point. For 99.9% of the population, those keywords
> indicate spam. For the 0.1% playing KOTOR, it does not. Hence, the ISP
> has to process each mail message just in case.
>
Actually, like most of the "false positives" the criteria going on isn't 
just keywords, its in certain combinations, with some crazy assed 
heuristic, which results it in not being so much as flagged "spam" as, 
"We detected, for no apparent reason, that this might be a threat, so we 
won't even show you the plain text, you have to explicitly say you want 
to see **the whole thing**." Umm, OK... But then, in other cases, you 
let me see enough of the plain text to see whether or not you falsely 
marked it, then let me tell you if its spam, or not. So, why the hell 
the difference?

In other words, "Possible real spam = we will let you tell us if it was 
or not", but, "Possible, non-existent threat = we won't even let you see 
it, until you decide to risk what ever threat we imagined existed, and 
then, if it isn't one, we won't let you tell us to stop doing it, over 
and over again, like we would with mere spam." :head-desk:

Makes no damn sense to me. If it wasn't a threat last time, how the hell 
is it next time, and why in bloody heck... Oh, wait, this is Microsoft, 
so they probably added their email equivalent of, "Are you sure you want 
program.exe to actually do anything?", to the bloody service... lol

Post a reply to this message