Chris Santerre wrote: >There are no guarantees in this world. And a chance at getting in the main >is about as good as Daniel baking me a cake. :)
At least Daniel is consistent :) >Again, the process usually goes: >1) I had idea, here is some code. >2) Code seems worthy, we shall test. >3) S/O is more wonderful then the invention of USB thunmbdrives. Yup. >4) *NOW* lets talk about -possibly- adding it to SA. Yup. >5) ???? >6) profit > >You are skipping to number 6 ;) Hardly. Planning is something businesses do. >(Alert, for the funny impaired, the above was a joke.) :) >>>I see alot of talk to support what is now vaporware that will help the >>>community by 'an order of magnitude'. >> >> >>I am have data which tells me it is much better than Razor. >>That I know with reasonable (statistical) confidence. What >>could be the boost to SA is unknown, although a boost from >>95+% to 99.5+% (5% to 0.5% fnr) would be "an order of magnitude". > >Exactly, you am have. We don't. Note the following data is from Feb. 7, so is out-of-date. We only launched in Dec, so rapid flux. My account is up to 99.7% now. >Date: Mon, 07 Feb 2005 16:31:06 +0800 >To: [EMAIL PROTECTED], [EMAIL PROTECTED] >From: Shelby Moore <[EMAIL PROTECTED]> >Subject: Significant evidence of AccuSpam's 98+% FNR > >Several high volume user accounts (% total spam correctly classified): > >Note in comparison I have seen "standard" (no training) SpamAssassin reported >in a >reputable study (for a single test user) at 92.5%. > >99.13% for 1840 spams and 233 non-spams (Shelby's email account) >99.86% for 2844 spams and 41 non-spams >98.96% for 3354 spams and 66 non-spams >99.21% for 3786 spams and 93 non-spams >99.71% for 2070 spams and 43 non-spams >99.11% for 1807 spams and 30 non-spams >97.92% for 1301 spams and 81 non-spams >99.17% for 727 spams and 13 non-spams >93.92% for 1250 spams and 166 non-spams >98.42% for 317 spams and 111 non-spams >95.79% for 451 spams and 110 non-spams >99.28% for 2095 spams and 243 non-spams >99.20% for 1741 spams and 29 non-spams >99.47% for 187 spams and 18 non-spams >100% for 1717 spams and 24 non-spams >98.09% for 943 spams and 18 non-spams >98.43% for 3384 spams and 92 non-spams >99.78% for 892 spams and 70 non-spams >100% for 1545 spams and 55 non-spams >98.82% for 850 spams and 27 non-spams >98.10% for 1314 spams and 105 non-spams >97.14% for 1538 spams and 23 non-spams >97.94% for 5618 spams and 72 non-spams >97.80% for 2779 spams and 382 non-spams >100% for 1885 spams and 22 non-spams >93.00% for 799 spams and 353 non-spams >96.67% for 751 spams and 29 non-spams >97.97% for 1529 spams and 33 non-spams >>If I sell services that include email and use Razor on that >>email, then I am selling "...commercial embedded software >>solutions...". > >Yes! If you are selling a service. SA doesn't sell anything. Although Theo >once tried to sell me a copy of the Pamela Lee.....nevermind. :) What are they going for? :) >Most people who submit to SA don't care where or how their stuff gets added. >Only that it helps the community. Yes, the word 'profit' doesn't sit well >with most of us. But you are in no way hampered by the fact that you want to >earn a buck. I assume most of you did not expend 1000+ man-hours on your contributions. Sharing the pool result of smaller contributions (open source) works well for some things, but not for all things.
