Paul Hoffman / IMC <[EMAIL PROTECTED]> writes:
> At 12:33 AM -0700 6/16/00, Russ Allbery wrote:
>> Brightmail is odd. I've never gotten a particularly good vibe from
>> them. I'm in general very leery of commercial filtering services
>> because in order to compete commercially they're pretty much out of
>> necessity keeping their precise techniques secret,
> They describe their techniques in detail on their web site. See
> <http://www.brightmail.com/isp/anti-spam/how/>.
I beg to differ. Their web site says, essentially, that they look at spam
and write rules to block it. Well, yes. Duh. What rules do they write?
Can I get a full list of all of their current rules?
I've not asked them for this, so perhaps I'm wrong, but it would really
surprise me if they were willing to post such a list on their web site. I
can get all the rules used by, say, MAPS.
> Brightmail puts all trapped spam in an easily-accessible folder so that
> the user (an end user or a sysadmin) can go through and look for false
> positives.
Yes, all the commercial spam filters do this (to my knowledge). But I
think this is fundamentally a bit of a cop-out; if I have to sort through
all the trapped spam, I may as well just get all the spam in my mailbox in
the first place. :)
If there's any real utility to the filtering, then stuff that gets
filtered out is going to be "less seen" in some fashion than stuff that
isn't, which raises the possibility of false positives. But then, I
personally am pretty paranoid about false positives; I know that a lot of
average computer users seem to routinely put up with levels of e-mail loss
that I'd consider completely unacceptable.
--
Russ Allbery ([EMAIL PROTECTED]) <http://www.eyrie.org/~eagle/>
