I think what Len is stating by saying Not concerned about FPS? > >It's simply not a problem, in practice.
Is that fact that if an IMGate machine uses the ORDB test to block mail (or RFC Ignorant, etc), 1 test, not a weight, it isn't a false positive. The sender is actually in ORDB (whether they know it or not, and whether the e-mail attempted to be sent was legit or not). In my eyes that is not a false positive. Sure they may be upset that their e-mail didn't get delivered, but they should direct that anger at the sending mail servers admin for not knowing what they are doing (most cases). <rant> I would have to say the number one problem with many mail systems is the copier boy is entrusted to set it up. The company thinks they are saving tons of money by having a paper certified moron do the installation. A little time goes by and they start receiving the so called FPs from Properly configured mail systems, and they go postal on those admins, not themselves for not getting their house in order first. </Rant> Jason ----- Original Message ----- From: "R. Scott Perry" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Friday, August 15, 2003 9:14 AM Subject: Re: [IMail Forum] Anyone have any experiences with Postini ? > > >>Not concerned about FPS? > > > >It's simply not a problem, in practice. > > So it's not a problem if free Internet services go away, and if you have to > pay higher prices from small businesses? See http://www.declude.com/plea.htm . > > >Religious point: content-scanners require 100% of all spam to be > >received/processed/stored. > > Len, please read before you post. :) David never said that content > scanning had to be involved! > > The problem is blocking on a single test that by design has false positives. > > >Here's a Wednesday's IMGate report for a very satisfied IMGate > >single-criterion heathen who runs three IMGates as equal-preference MX. > > But, that "very satisfied" person does not know how many false positives > they are getting. Worse, they have no idea of the cost of the false > positives (since only part of the cost of the false positives is on their end). > > > 169924 rejected (84%) <<<<<<<<<<<<<<< > > That's cute. But if you use the URBL test (see > http://www.declude.com/junkmail/support/ip4r.htm ) you'll reject 100% of > your E-mail. So are you saying that you aren't catching nearly as much > spam as you should? Of course not. Your 84% figure is meaningless. > > Wasn't there a book called "Lying through statistics"? Your statistic -- > rejecting 84% of E-mail -- is meaningless in almost all spam > contexts. Heck, if you want statistics, in July, Brightmail claimed 50% of > E-mail was spam. 84% minus 50% is 34%, so if their stats apply to your > system, so you are rejecting some 34% of your legitimate E-mail. Let's not > play games with stats (for those reading this, I'm sure Len's FP ratio is > much less than 34%, although it wouldn't surprise me if it was as high as > 5% to 10%). > > >Here's the key point: those 32K deliveries are 1 GB of volume. Let's > >assume from those delivery numbers that the 160K single-criterion rejects > >(rejected after RCPT TO: so we never accepted the msg DATA, the reject has > >minimum cost) represent 5 GB of traffic. > > So you are rejecting 5GB of traffic. So what? If you stop your poor > "block-mail-on-a-single-spam-test" philosophy, and change it to > "block-mail-that-fails-several-spam-tests", I'll bet you could block 4.5GB > of traffic -- and that some .4GB of the .5GB that you are no longer > blocking would be extra legitimate E-mail. > > All done without smoke and mirrors -- no high CPU content filtering, no > accepting 5GB of mail. > > >What you're recommeding is that this IMGate admin would be much better off > >sucking in those 5 GB... > > Come on, Len. Read the E-mail before you respond. You're so stuck in your > ways that you are refusing to even see that there are other options. Have > you even bothered to notice that *my* suggestion can't even be done with > Declude? I'm not doing this to argue with you, or to sell my product, or > because I'm having a religious experience. I'm doing it so you can improve > your tool! > > Just admit that you can't use a weighting system on your platform, and be > done with it. Or add one. Or explain (with facts, please) why a weighting > system wouldn't help significantly reduce false positives. > > >>Weight testing allows recipients to make their own judgement as to what > >>is legitimate email as oppsoed to spam > > > >IMGate admins are fully competent, and successful, to make their own > >judgement as to what they allow into their systems. > > Not when you give them FP ratios that are made up, or when you urge them to > do something that blocks 5% of their legitimate E-mail without explaining > to them the drawbacks. People trust you Len. So their judgement isn't > "I'm willing to block 5% of my legitimate E-mail", it's "Len says this test > works wonders, and I trust him, so I'm going to use it." > > If you want to mislead people (either intentionally or not), I'm not going > to stop you. But if you do so here, I'm going to respond, so that people > will know exactly what is going on. > > >>but only in a very narrow range of circumstances. > > > >Wrong. A single criterion is sufficient to reject huge volumes of > >crap. see above report. > > Yes. Blocking any E-mail coming from an E-mail address with ".com" in it > will reject huge volumes of spam. But you're ignoring false positives. A > single criterion is only useful if it has a very low FP ratio. > > >>Defeats the object of email in the first place. > > > >Nope. What defeats the object of email is having spam volume running at > >50% to 90% of total mail volume. ( above, precisely 84% for Wednesday, 24 > >hours) > > Geez Len, stop misleading people. The spam volume on that server is NOT > 84%. That's how much E-mail you are blocking. The actual > spam-to-legitimate-mail ratio is probably about 80%. > > -Scott > --- > Declude JunkMail: The advanced anti-spam solution for IMail mailservers. > Declude Virus: Catches known viruses and is the leader in mailserver > vulnerability detection. > Find out what you have been missing: Ask for a free 30-day evaluation. > > --- > [This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)] > > > To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html > List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/ > Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/ > To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/ Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/
