On Tue, May 11, 2004 at 09:47:14AM +1000, Matthew Palmer scribbled: > On Mon, May 10, 2004 at 02:44:38PM +0200, Marek Habersack wrote: > > On Mon, May 10, 2004 at 01:08:02PM +1000, Matthew Palmer scribbled: > > > On Mon, May 10, 2004 at 04:14:28AM +0200, Marek Habersack wrote: > > > > How about not using spamassassin instead? > > > > http://www.nuclearelephant.com/projects/dspam/ - a much faster and accurate > > > > tool... > > > > > > And, naturally, you have hard numbers to back up your claim, for situations > > > analogous to the Debian mailing lists. Because it would be terrible to make > > > sweeping statements like that with no evidence. > > One more thing. It wasn't a statement, rather a thesis which may or may not > > be confirmed - but that requires testing. And something whispers in my ear > > "A much faster and accurate tool" seems like a claim requiring some evidence > behind it to me. It wasn't "it's faster than SA and might be worth a try", > or anything like that -- you stated, categorically, that dspam was faster > and more accurate. For that, I would expect anyone to be able to back up > their claim with some evidence. Otherwise you're just talking out your > arse. I expect some intelligence from the audience, sorry for that. I strongly believe in concise communication and brevity. And that dspam is faster is obvious for anybody that has ever written a single line of code. > > that spending some time testing a possibly better solution is better than > > spending $$$ on new hardware. So, please refrain from sarcasm for the future > > and if you don't have anything substantial to say, don't say anything. There > > I had something quite substantial to say. I asked for something to back up > your claim that dspam was faster and more accurate. You've got nothing, If you spent a little of your time investigating instead of writing useless and pointless mails, then you would have found the following on the DSPAM page: A strong focus on large-scale implementation support. The largest implementation of DSPAM we've heard about to-date involves 125,000 users. DSPAM has been designed to experience a very short execution time (between 0.01s - 0.03s real time for classification and between 0.03s - 0.10s real time for training, on average hardware), and has been equipped with a storage driver API allowing several different storage mechanisms to be used. Depending on disk space constraints, accuracy can be traded off for additional disk space or vice-versa. Note the numbers with little 's' next to them? It's the time. Now, SpamAssassin doesn't go below 2s on average. Is that faster enough for you? > apparently, because you need help to implement a test regime. Now, that's And, I understand, you offer that help? I will gladly learn from the wise ones. > fine -- test away, and if it turns out to be good[1], you'll see me at the > front of the queue asking politely for it to be implemented. But you can't I couldn't care less for your asking. If I test it, I will do it for myself and share it with those who are interested in it. > come and make a bald assertion, with no evidence, and expect anyone sane to > pick up your banner and charge into battle on your behalf. Aren't you drawing just a tad bit too many conclusions, buddy? You certainly have a way of remote mind reading or supernatural ability to read between the lines. That's way cool, I envy you. > > is already enough of hostility, sarcasm, self-importance and similar nice > > things on the debian lists. Thanks. > > I'd vote "self-importance" for your initial post, if I were restricted to Your vote is counted. And I'm just sick and tired of the likes of you who wrote a smartass-style mail and contribute nothing to the subject whatever the subject is. > those choices. More accurate would be "uninformed hyperbole". Yep, my Also a nice expression. > response was sarcastic with maybe the tiniest dash of hostility, so it's a > good thing this isn't the touchy-feely brigade. Perhaps next time you'll > come up with something, anything, to back your claim up with, which will > mean there'll be one less dork thinking that proof by assertion is the way > to get other volunteers to do the work. Dear boy, once again - you should have spent at least a minute of your time doing some research on your own. Call me naive, but I tend to believe what the author of the software wrote about it - and I have _every_ right to do so and repeat his claims wherever and whenever I want. If you claim that I say bullshit (and all that nice stuff you've said above), then you should repeat that in a mail to the DSPAM author - because you're saying that HIS words and claims are bullshit. > > [1] My personal list of "good" things is: has to be non-resource-intensive, > have lower false negatives and at least as low a false positive rate, and > require minimal initial and on-going manual maintenance. I'm sure the > listmasters have a couple more requirements -- you should ask them. Why would I? You've lectured me already. Unlike you, I won't waste other people's time. Now, before you go on with your lecture, go ahead and read http://www.nuclearelephant.com/projects/dspam/ - you might find some answers to your questions. And one more thing - next time you claim somebody is 'uninformed' or 'makes bold assertions, with no evidence' etc. - be prepared to prove him/her wrong because your statement is as uninformed, as bold as mine was. I didn't show numbers (assuming that if somebody speaks up on the subject they would have read the DSPAM page and found the numbers given there) but neither did you. So please, kindly refrain from further commenting unless you have something really substantial to say - I don't give a damn about your opinion about me, so spare yourself some time and read a book or something. marek
Attachment:
signature.asc
Description: Digital signature