http://bugzilla.spamassassin.org/show_bug.cgi?id=1375
------- Additional Comments From [EMAIL PROTECTED] 2004-03-02 15:57 ------- Re: Dan Hollis' comments Whatever disagreements may be left, I think we are very close on what we are proposing. You say we should skip hrefs that are invisible due to having empty link areas. I say we should skip them and as long as we are looking for that also look for ones that match the existing small font and invisible color tests. The purpose is the same as testing for empty link areas and we already have the code. I did not propose looking up image references to decide if they are visible: I mentioned that as something which would _not_ be practical to do. I agree that we should implement rules that will help even if they are not perfect, which means that we should have this DNSBL rule even if we cannot tell what an image only link points to. I agree that if spammers have to include hundreds of links to obfuscate one real one, that in itself would be a good trigger rule. The point I was trying to make is that if we include the DNSBL rule we should 1) use a small random sample to avoid DoS, and 2) Add a rule to catch too many links. So really, we are in agreement, but I'm extending your suggestion that we use the DNSRBL rule modified to ignore empty link text. The extensions are 1) Ignore for the DNSBL rule not just empty link text but anything that matches the existing tets for "invisible" text based on font size and color 2) Count the number of such ignored links for a possible separate rule that penalizes them. There's not much point for invisible links in real email. 3) Separately count the number of links with an image-only clickable area and score a rule for that separately. Assume that they are not invisible for the purpose of DNSBL checking, since invisible ones will only be useful to the spammer in quantity and the quantity will trigger this rule. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
