Comments on: Well, that was fast http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/ because the ass won't kick itself Sat, 05 Jul 2008 19:38:18 +0000 http://wordpress.org/?v=2.5.1 By: Ryan http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-5204 Ryan Wed, 19 Apr 2006 04:20:36 +0000 http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-5204 Yeah, the thing is, back when I was doing this I was hopping from one anti-spam solution to the other. So it was more important to keep that corpus of data in case I opted to switch to some locally-based filter. Yeah, the thing is, back when I was doing this I was hopping from one anti-spam solution to the other. So it was more important to keep that corpus of data in case I opted to switch to some locally-based filter.

]]> By: FiReaNG3L http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-5199 FiReaNG3L Tue, 18 Apr 2006 20:47:28 +0000 http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-5199 Actually I think that Akismet is bayesian-filter based, so keeping you old spam to train yourself a filter is kind of pointless (as long as their service stay up, anyway). But what happen if they go down / out of business? Actually I think that Akismet is bayesian-filter based, so keeping you old spam to train yourself a filter is kind of pointless (as long as their service stay up, anyway). But what happen if they go down / out of business?

]]> By: Ryan http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3260 Ryan Wed, 26 Oct 2005 18:31:42 +0000 http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3260 Ha...actually the only reason I've kept all my spam around is because I've hoped that there might be some sort of bayesian spam plugin at some point, so the corpus of spam would actually be useful. And Akismet didn't flag your post because I currently have it disabled. I'll be reenabling it tonight to give it a second go. Ha…actually the only reason I’ve kept all my spam around is because I’ve hoped that there might be some sort of bayesian spam plugin at some point, so the corpus of spam would actually be useful.

And Akismet didn’t flag your post because I currently have it disabled. I’ll be reenabling it tonight to give it a second go.

]]> By: IO ERROR http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3259 IO ERROR Wed, 26 Oct 2005 12:27:49 +0000 http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3259 You're a strange one. Most people don't want to <em>keep</em> their spam around. :) Are you secretly a fan of texas hold 'em and cheap home mortgages? No? How about phentermine and viagra? Now, see, Akismet didn't flag this post as spam, despite all the nasty keywords, so it's a little smarter than you think. :) Most of us testing Akismet are seeing a <a href="http://error.wordpress.com/2005/10/26/automattic-kismet/" rel="nofollow">0.1% false positive rate</a> and that rate is dropping. That's easily better than most other spam plugins out there. Now, why won't you come to my online gambling site? :) You’re a strange one. Most people don’t want to keep their spam around. :) Are you secretly a fan of texas hold ‘em and cheap home mortgages? No? How about phentermine and viagra? Now, see, Akismet didn’t flag this post as spam, despite all the nasty keywords, so it’s a little smarter than you think. :)

Most of us testing Akismet are seeing a 0.1% false positive rate and that rate is dropping. That’s easily better than most other spam plugins out there.

Now, why won’t you come to my online gambling site? :)

]]> By: Ryan http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3258 Ryan Wed, 26 Oct 2005 05:46:56 +0000 http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3258 Yes, that's exactly what happened. The 440 spam that showed up were ones I had previously marked as spam dating back to May of this year. I was shocked to see that it had instantly flagged 440 pieces of spam until I finally realized what was going on. As for the 15 day thing, that totally makes sense. It gives you time to play. That being said, you know on about day 10 that there's a clock ticking in your system that you have to constantly be aware of if you don't want to lose anything in your system. Maybe when I get some time tomorrow I'll edit the PHP file and comment out the call to remove the old spam and then let the system run a while. I think I ought to be able to change all of the "spam" in my database to something like "oldspam" so Akismet doesn't see it anymore. That would at least let me start with a clean slate with Akismet while still retaining my old spam. Yes, that’s exactly what happened. The 440 spam that showed up were ones I had previously marked as spam dating back to May of this year. I was shocked to see that it had instantly flagged 440 pieces of spam until I finally realized what was going on.

As for the 15 day thing, that totally makes sense. It gives you time to play. That being said, you know on about day 10 that there’s a clock ticking in your system that you have to constantly be aware of if you don’t want to lose anything in your system.

Maybe when I get some time tomorrow I’ll edit the PHP file and comment out the call to remove the old spam and then let the system run a while. I think I ought to be able to change all of the “spam” in my database to something like “oldspam” so Akismet doesn’t see it anymore. That would at least let me start with a clean slate with Akismet while still retaining my old spam.

]]> By: Matt http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3257 Matt Wed, 26 Oct 2005 05:25:31 +0000 http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3257 That makes sense. I guess from my point of view the 15 days is enough time to review what it's caught (you can mark things as "not spam" too and it learns from that) and I'm pretty resistant to adding options on principle. Perhaps the key misunderstanding was the 440 comments were things that were already caught by WordPress before Akismet got there, probably from words in your blacklist or things you've marked as spam before, they've just been invisible until the Akismet UI exposed. I'll give some thought to how to make that easier. That makes sense. I guess from my point of view the 15 days is enough time to review what it’s caught (you can mark things as “not spam” too and it learns from that) and I’m pretty resistant to adding options on principle.

Perhaps the key misunderstanding was the 440 comments were things that were already caught by WordPress before Akismet got there, probably from words in your blacklist or things you’ve marked as spam before, they’ve just been invisible until the Akismet UI exposed. I’ll give some thought to how to make that easier.

]]> By: Ryan http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3256 Ryan Wed, 26 Oct 2005 05:20:25 +0000 http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3256 I don't actually mind Akismet wiping out the comments, but I want it to be optional for a little while so I can test it out without fear of losing anything. The problem is that I don't know Akismet yet so I don't trust it. I'd like it to not mess with anything until I've reached a level of trust with the system. Sort of a trial period where I could see what it would flag before deciding that it's not going to cause too many problems. Thanks for responding. Maybe I'll hack up the Akismet plugin a little bit to get what I want. I really do want it to be good, god knows I need the help fighting off these spammy bastards. I don’t actually mind Akismet wiping out the comments, but I want it to be optional for a little while so I can test it out without fear of losing anything. The problem is that I don’t know Akismet yet so I don’t trust it. I’d like it to not mess with anything until I’ve reached a level of trust with the system. Sort of a trial period where I could see what it would flag before deciding that it’s not going to cause too many problems.

Thanks for responding. Maybe I’ll hack up the Akismet plugin a little bit to get what I want. I really do want it to be good, god knows I need the help fighting off these spammy bastards.

]]> By: Matt http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3255 Matt Wed, 26 Oct 2005 04:00:07 +0000 http://unclehulka.com/ryan/blog/archives/2005/10/25/well-that-was-fast/#comment-3255 I was saving my spam comments for "historical reasons" as well, but then it started building up and one day I looked and my wp_comments table was over 15MB. We were also getting a ton of complaints about bloated backup files. The original reason was for some sort of analysis of the spam for plugins, but it's been over a year since the feature was added and no one has does that yet. However Akismet does do that, just in a centralized fashion and on a global scale. However if you really want to save the spam, look for the akismet_delete_old() function, deactivate that and you're set. I was saving my spam comments for “historical reasons” as well, but then it started building up and one day I looked and my wp_comments table was over 15MB. We were also getting a ton of complaints about bloated backup files.

The original reason was for some sort of analysis of the spam for plugins, but it’s been over a year since the feature was added and no one has does that yet. However Akismet does do that, just in a centralized fashion and on a global scale.

However if you really want to save the spam, look for the akismet_delete_old() function, deactivate that and you’re set.

]]>