|
Per Theo’s suggestion, I’m submitting my offer to contribute my open-source challenge-response Spam solution to the SpamAssassin project. Would this be of interest to the group?
Thanks! Brad
On Wed, Sep 29, 2004 at 09:21:05PM -0600, Bradley D. Brown wrote: As Bob suggested, I would like to contribute SpamFree as an open-source option for SpamAssassin. How do I go about doing that?
Hi,
I'm glad to hear you wish to contribute, thanks! :)
The best place to go is [email protected]; it's the list that all development issues are discussed on, so it'll hit everyone it needs to. :)
-----Original Message-----
Theo,
As you can see below, I was posting to the wrong list server…I was hoping to figure out who I needed to talk to about this. As Bob suggested, I would like to contribute SpamFree as an open-source option for SpamAssassin. How do I go about doing that?
Brad
==========================
BobMenschel
> 99% is certainly respectable. My thought was that it would be nice to close the final 1%.
We're actually running between 99.5% and 99.8%. I agree it'd be good to close those final few, but IMO the only way to do that without false positives is to have mind-reading StarTrek-level AI software.
> My thinking is that this would be another option that people could choose to turn on...or not.
Then you're posting to the wrong forum -- here we only deal with simple rule add-ons for SpamAssassin, and hints on the use of SpamAssassin. We're too far removed from the real action, which is what you're talking about.
Your best contact is probably with the SpamAssassin development team. SA web page is at http://spamassassin.apache.org/. Theo Van Dinter might be the best one to contact if you're interested in joining the SpamAssassin project and contributing SpamFree as an open-source option to be used within SpamAssassin.
Brownb
99% is certainly respectable. My thought was that it would be nice to close the final 1%. We have spam filtering software in place and it also provides some false positives. SpamFree "fixes" the false positives too. My thinking is that this would be another option that people could choose to turn on...or not.
As for your question, yes, if user A doesn't pre-add user B, user A will receive a message asking to confirm they aren't a spammer. However, there are a number of ways to avoid this. 1) pre-add user B to SpamFree before sending the message - this is a simple email message to SpamFree, 2) pre-add user B's domain (i.e. if this is a customer or someone that you want to receive everyone in that company's email - it then does spoof checking), 3) Include a magic work in your subject (one that's been pre-added as a wildcard subject). The final option is to wait until user A receives their daily Spam listing email which will include the message from user B. The link to add user B is in the email. The daily email also groups the message by confirmed spammers (i.e. the message to them bounced) and still unknown/waiting for response, which are ordered by those making it through the spam filter first. In other words, on a daily basis, out of the 700+ spams I receive the messages that meet the example you provided are typically right at the top of the list. Another option would of course be to add functionality to SpamFree to allow message to be cced or bcced to SpamFree, which would automatically add user B to your friends.
BobMenschel
I'm not sure that our activity would be the place for your software, since we focus almost exclusively on SpamAssassin, and thanks to SpamAssassin our users are 99%+ free of spam (unlike the stats you suggest in your marketing presentation).
However, I'm curious about how users of your software use it. Hypothetical: User A of your software sends a message to some person B somewhere on the Internet. Person B responds. Does his response get blocked by your software?
Personally, if I sent a message on my initiative to your User A, I wouldn't mind doing the SpamFree validation to get my message to him.
But if I took the time to answer your user A's query, and that query is blocked by your software, under his command, I'm not going to waste my time jumping through hoops to get that response to him. It's up to your user, that requested the response, to open the door.
My experience is that the majority of users A don't think ahead at that level, and if others have the attitude I do, they lose a lot of responses.
Does your software have some method of automating the "allow /responses/ through" challenge?
Brownb I wrote an open-source challenge-response package (SpamFree) that I'd like to contribute to this project. Is that possible? If so, how would I go about doing this? Who do I need to talk to?
More information about SpamFree can be found at:
http://www.tusc.com/oracle/technology/tech-spamfree_1.html SpamFree
Thanks for any help that you can provide. Brad
Bradley D. Brown
|
