-----Original Message----- From: Jeff Chan [mailto:jeffc@surbl.org] Sent: Saturday, October 09, 2004 9:06 PM To: SURBL Discussion list Subject: Re: [SURBL-Discuss] Revised DMOZ data, got Wikipedia domains too
On Saturday, October 9, 2004, 1:09:43 PM, Patrik Nilsson wrote:
Just create a separate "TLDs or treat as TLDSs" zone that
can be checked
and cached client side.
Or even better - give "TLDs or treat as TLDSs" a
distinguished A value in
existing lists. If a lookup returns XXX.XXX.XXX.XXX, it is a "TLD or treat
as TLD" and
should be further recursed. If we think there is a risk that some bad client
implementatins treat any
returned A record as a hit, use TXT records.
This is still an interesting idea, but I'd still be somewhat concerned about putting out a list that looks like a regular SURBL that it could get misused.
But perhaps the larger issues is that the hard core spammers don't seem to use *subdomains of legitimate shared-domain hosting providers*. They just register their own full domain names and use those (lots of them).
If some legitimate hosting provider has an abuse issue, then it's in their own interest to stop the abuse.
SURBLs are arguably best suited for cases where the ISP is spam-friendly and allows spam hosting on custom domains. The reality is that's a much larger and tougher problem than shared, common-domain hosting, like a geocities or tripod.
The best use of our time is to focus on the biggest spammers first, and we're not catching all of those yet.
Big surprise...I disagree :-)
I think we are already reviewing these domains, so why not just add to catch the smaller ones instead of throwing them away.
But I see your vision, and I can bend like the reed ;)
But given enough time, and perhaps enough chocolate, I think I can turn you around :-)
--Chris