http://bugzilla.spamassassin.org/show_bug.cgi?id=1375
------- Additional Comments From [EMAIL PROTECTED] 2004-03-02 15:21 ------- Subject: Re: do RBL look-ups on URLs > I see one major problem with skipping invisible links in at least some > form of examination. That is where the spammer will put the web bugs. > After all, they don't need to be (and generally aren't) visible. Web bugs are usually URLs that are loaded without any action on the part of the user. We can always test those. > I consider the "lookup DOS" argument somewhat moot. Count the number > of links in the message and divide into the text size. If there are > more than x*10^n links on the page, or the links are more than y% of > the total body size, declare the thing to be spam without looking > anything up. We can try that as a separate test. A test for "more than y%" probably won't work (even with a minimum length, I suspect) since non-spammers do that all the time. > Incidentally, this makes an argument for self-scoring tests. A test > that counted URIs in the body and gave a score based on the number of > hits (possibly times a factor or with upper and lower bounds) would > make this sort of decision a lot easier than it currently is. We do that with range tests all the time. Breaking tests into ranges is a bit clumsy and I think having a scoring function would work better, but we lack the code to handle it. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
