Subscribed accounts shouldn't be used as spamtraps, and they should work harder at removing impure spamtraps with their system. I also think it would be a very good idea for them to come up with less and more spammy versions of their database, which would allow you to reject on the spammier, and score lower on the less spammy, thinking that things like Ebay, PayPal, Match.com, etc. will still get through.
Also, to think that they will take just about any submission and weight it without an understanding of the individual's own standards is flawed. Just like they score spam, they should also score users and spamtraps. Seriously, if you are a person that reports PayPal when you are their customer, the weight of your opinion should be lower. I think that this could probably be done in an automated fashion and it would improve their system immensely. Right now, virtually any mail-blast from a large company finds their way on the list or at least teeders with the limit.
Another problem is that over time, the quality of their participants goes down. This is based on the fact that formerly it was only the best informed that knew about the system, and now they are much better known with less knowledgeable participants. I don't think that maybe the added value of even more participants now detracts from the overall value enough that it should be limited. I'm of course speculating here, but it seems logical to me.
Matt
Colbeck, Andrew wrote:
Well, it's important to remember that SpamCop is user-driven.
The man behind it, Julian Haight, and his Spam Cop deputies focus on parsing the messages well, holding off the DoS attacks, juggling the expiry and the weight of the IP & subnet based on reports, and getting the right abuse addresses and that's about it. Who gets listed really isn't their deal.
In that way, it's a lot like CloudMark, only it doesn't have a counterweight system. The only safety valve is the expiry time, or users like us complaining in their newsgroup about unwarranted listings.
As Chuck says, it simply can't be used by itself reliably. As for AOL mail, I think they've come a long way. I used to counterweight mail from .mx.aol.com to counteract the IPNOTINMX and NOABUSE and NOPOSTMASTER weights it would always fail, and I recently found that they've gotten much better at containing spam; they still host reply mailboxes, but are sending out very little to us, so I've increased my counterweight for mail coming from their mail servers.
Andrew 8)
-----Original Message-----
From: Matthew Bramble [mailto:[EMAIL PROTECTED] Sent: Wednesday, September 24, 2003 8:12 AM
To: [EMAIL PROTECTED]
Subject: [Declude.JunkMail] What's wrong with SpamCop?
Is it me, or did SpamCop suddenly become awful when it comes to false positives with almost anything that is sent in bulk? I've recently seen them tag PayPal, ActivePDF newsletters, Match.com and even the local chamber of commerce (which only sends to members w/opt-out). If they ever start crossing FP's with MailPolice, two very important RBL's will suddenly become greatly diminished in value on my server.
So the question is, does SpamCop care about this problem? Are they going to make fundamental changes in how they determine what to block based on their clearly impure input? Anyone have a scoop?
Matt
--- [This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]
--- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
