seems like new antispam services are en vogue ;-)
anyone tested this one? http://www.defensio.com/
seems like new antispam services are en vogue ;-)
anyone tested this one? http://www.defensio.com/
I have a question about identifying "unidentified" spam... is this done per user, or by the super admin? What about the bayesian scores themselves, is this per-blog or global? Thanks for the info, looks like an interesting plugin...
Identifying "unidentified" spam is done by only the Super Admin. No doubt you're thinking along the lines I was: don't want to let the spammer be the judge of what is Spam. A future version may give users to flag misidentified spam, but only to the extent that it informs the Super Admin that this message needs to be checked. So far, misclassification doesn't seem to be an issue after an initial training.
Bayesian is global and so far seems to be fairly robust. Testing will tell. I'm just about to send the first round of Beta testing to people (a power cut in our building hasn't helped my sanity this week :-) so if people haven't heard from me on this forum they will soon. (I want to test installation with a few select users first to get past that but a general Beta test).
Thanks for the reply macgruder, actually I am in the early stages of setting up a Mu based site where the users will be paying for accounts (there is much more to the site than just the blog) so splogs aren't a problem for me and I'd probably want users deciding what's spam. Although if it works as well as you're saying then I suppose the volume would be low enough for me to handle it.
Is this designed to work in conjunction w/ Akismet? (Maybe by giving extra "spam points" to comments that Akismet catches...)
Comment spam is the really the hardest to moderate because it is not a particular user. Even if splogs are not a problem, you will eventually be hit with hundreds or even thousands of comment spams a day.
The plugin doesn't work with Akismet because it's not necessary to be honest. A bayesian filter recognizes the patterns of the users who are members of the site in question and adding an external element won't improve the accuracy (and may even reduce it), but would add more work for the site owner to check.
Much of the strength of Bayesian filter comes from their focus and smaller sample size. Generally, Bayesian filter tend to be extremely robust. Issues like spam poisoning tend to help them become more accurate rather than less.
anything new on this?
Yes, apologies all round. We had a major powerout in our whole building that put my website out of action for 3 days (ouch)! I didn't feel up to dealing with Beta feedback during that period of stress and sleepless nights :-)
Here's a screen just to let people know how it's going :
I am interested in testing it, I am setting up two blogging sites at the moment, and this is my biggest concern, would be amazing if it works like you say it does!
could it have a problem with other languages?
email: cabejon [at] gmail.com
Hi !!
I'm interested too, if it still open.
matbasili for gmail dot com
Thanks !!
Same thing than previous message.
thierry / 2diabolos * com
Thanks a lot.
I would like to say that for a beta release in August, there have been some delayed ...
sorry
Hi.
I am also interested in testing this tool.. My Blogging system is full of spam. :(
I could test it 4 u. Mail: tomas(at)srna(dot)sk
Thanks
interested in testing too :)
i've just installed the latest mu version 1.3 and can stand a little testing. so if you want to know if it will work on 1.3, let me know. and, in return, a dutch translation maybe?
let me know via ravanhagen - gmail - com
--ravan
Any updates on this?
seems like its gonna take soem more time, in the meantime check out this new antispam service: http://defensio.com here is a thread I started about wpmu compatibility; http://forum.defensio.com/comments.php?DiscussionID=10
Managed to get back to work on this. Sorry about the delay. I should be sending out the first beta's imminently - hopefully this weekend. (apologies to those who have heard it before!). I discovered through my own beta that certain administrative tasks didn't work as expected requiring a bit of a rewrite.
Well, I could help you test it ;) mijaeus [] hotmail.com
OK. I've sent out the first round of Beta's .
If you've posted here, and haven't heard anything, don't worry, I'm going to wait for the first feedback, just in case there are obvious errors or problems.
Macgruder.
OK. I've sent out messages to everyone who posted here. If you didn't get the message, please check your spam filter (!)
Then we can try a real test to finally defeat the bastards :-)
Macgruder
Any updates? Curious... I've been slowly eradicating spam comments, but it's the splogs that are killing me at the moment.
Yes, I've been keeping it up to date. It stops splogs fairly well for me, but I don't have a major splog problem so I'd be interested in testing it out on someone who does. I also added trackbacks.
One major addition is that it tracks the IPs of people who spam and attempt to spam and when they reach a certain thresh-hold they are blocked instantly. This saves a large amount of bandwidth and processing power. For example, I have one IP that has been stopped over 1000 times.
You can mail me at
monty_2008 /at/ j hyphen w3 d0t com
just to give some feedback:
I have been using it since the first beta round and am quite happy with it, although, I only have a testing install, with just 1-2 real bloggers so my tests are not really that representative.
Just keep mcgruder up to date with feedback.
I would love to test this, especially when I have soooo many spam blog posts right now. Add me to the list please!
Fryer
A sudden renewal of interest the day before I go on a 10 day trip until March 7th.
To everyone who is interested:
Send a mail to me at
monty_2008 /at/ j hyphen w3 d0t com
hyphen means '-' without the quotes!
with the DOMAIN(S) that you intend to use it one. Just the top level one is fine. I will send everyone stuff when I get back.
My stats are as follows (2 months):
Comment Spam stopped: 15281
Mistakes: 3
Track Spam stopped: 153
Mistakes: 1
Stopped Dead: 21505
The stopped dead is significant because these represent consistent spam IP addresses, and so they are blocked immediately without using up major resources on your server. Since they don't get a chance to post or attempt to signup, you can't know which they were attempting to do which is why perhaps I get no splogs.
I'll be contacting people shortly but yesterday
I broke my arm
Tripped over Monty. One of the dangers of working with a python.
He was like, I sure as hell can program this stuff, but who's going to do the typing?
Out of cast in a 10 days or so.
I hope you recover soon!
:-)
now you have a real reason to type 1-handed. LOL
your python is named Monty? I dunno whether to laugh or groan. ;)
Get well soon bud.
Will a tester send me a copy? I sent an E-mail to monty, but I am getting attacked by those little spammer errrghs...... right now.
bear /at/ political bear /dot/ com
whats the update on this spam blocker?