On Fri, 2008-03-28 at 11:06 -0700, Craig White wrote: > On Fri, 2008-03-28 at 10:17 -0700, Les wrote: > > On Fri, 2008-03-28 at 09:35 -0700, Craig White wrote: > > > On Fri, 2008-03-28 at 08:29 -0700, Les wrote: > > > > > > > > > > But is 99.99% delivery sufficient? I receive more than 150 emails per > > > > day (ones that I am interested in), and every few days I need to receive > > > > certain emails about customer relations and ongoing projects. 99.99 > > > > percent means I would miss one every 66 days. If the one that I miss > > > > cost me a contract, it might not matter whether I received the rest or > > > > not. Currently I have to parse through the junk mail locally and > > > > remotely about once a week. the ISP junk folder often has more than > > > > 1200 emails in it. THe local one about 30. This adds about 1/2 day of > > > > overhead every week to recapture what should have come through. > > > > Personally I think the world needs to eliminate spam, or at least make > > > > every effort to seriously reduce it. > > > ---- > > > but your example completely misses the point. > > > > > > the 'Junk' directory is a result of some type of agent parsing accepted > > > e-mail, scoring it and redirecting it based on a score. > > > > > > The point of greylisting is always about (or virtually always...depends > > > upon various implementations anyway) sender/recipient/smtp server > > > 'tuples' and 'Temporary Failure' /SMTP result codes 450 and whether the > > > sender attempts redelivery. > > > > > > http://www.greylisting.org/ > > > > > > I would suggest that rather than prove the argument about the problems > > > with greylisting, you have proven the opposite because if you can lop > > > approximately 70% of that junk mail off the top via greylisting, you > > > wouldn't have to look through so much 'junk' to find the false positives > > > that inevitably occur with any type of spam scoring system. > > > > > > Craig > > > > I cannot argue the value of greylisting. But efficiency is in the eye > > of the beholder. My time is limited. My work valuable, and my customer > > correspondence is dictated by my customer, not my email policy. I am at > > the mercy of the ISP here, and the customer. Yet I am the one who > > suffers loss, not the ISP, not the customer who will find someone to do > > his work. Yet you as the web person thinks this is effective. I cannot > > speak to school systems, only professional uses. Time is money and lost > > opportunity is even more valuable, resulting in loss that cannot be > > measured, yet ultimately may determine the success or failure of my > > business. > > > > I would like to take this thought "out of the box". The methods > > currently in place, grey listing, parsing for key words, and other > > simplistic means, while effective at reducing traffic are not really the > > desired solution by users, who would like 100% success in getting their > > desired email. And while I cannot spout statistics about loss, I know > > it happens from personal experience. The question at hand is how to > > avoid even more loss. It is a quality issue, and today the quality > > standard is not 99.99%, it is 99.9999% (six nines or six sigma) in most > > industries. To say that 99.99 is good enough is the path for GM and > > Ford, not Toyota and Nissan. Think of it that way. So how do we > > improve by two decades of quality in this "war" against spam? > ---- > greylisting is just one of a lot of tools available to the mail server > administrator - all of them are calibrated to minimize the junk mail > delivered to the end users and of course minimizing delivery failures > and false positive scoring by various mechanisms. > > End users of course are the ones ultimately affected and you seemingly > want to open an end user discussion about a server level > technology...please don't as it won't provide clarity to anyone. > > If you are losing e-mails, evaluate the filtering system that you use at > end user level and discuss the methodologies employed by your mail > provider with them. > This is the same sort of nonsense that the carmakers spouted about quality in the 70's. See where it got them. This is an end user product. Clarity is the way the user sees it, not the way the provider sees it. I know there is no sense arguing this, I have dealt with engineers and quality issues for years. The only way that attention comes is when someone provides the user with higher quality and the engineer is provided with a pink slip when his job goes away. It is coming. Wait for it... wait for it.... regards, Les H
