 |
 |
Be aware that the code was written years ago as an example of what COULD be done with client-side spam filtering and the new agent trigger, before new mail arrives. It was never an IRIS-supported agent (no sandbox code is supported....it is use at your own risk). It was never intended to be marketable, bullet-proof code. I have played with it lately to make the menu commands more like Outlook ("Actions - Junk Email"), run on command (as well as on schedule), load an array of spammers instead of looping through the list every time (better performance), and automatically remove duplicate spammers, but otherwise I haven't really done anything major to it. Once you publish an article in Iris Today, Iris gets the rights to the code. So technically I would not own the rights to create a marketable product or make updates.
One of the reasons that the agent and accompanying UI couldn't be more usable is that I didn't want to modify any of the existing design elements in mail. So there are definitely better ways to do things that I couldn't do because of that limitation (changing the mail template technically makes it unsupported by Lotus).
Of course you should be employing server-side spam filtering....R6 is much better than R5 because the server can subscribe to the blacklist. But of course this will always be out of date. And don't forget the other R6 filters (IP address, domain, etc.) to block incoming mail.
I have found many spammers send messages using my own email address/domain, so it is hard to filter that out! And of course, they keep making up new return addresses, so you really need to have some intelligence that also looks into subjects and bodies to compare against a banned word list. That sounds like an interesting project! The new R6 Configuration Rules attempt to do this but I have a hunch you will really kill server performance if you go wild here.
There are a couple commercial Domino add-ins that do spam filtering based on content. The only issue is cost, of course! You can also employ an SMTP filter that runs even before it gets to Domino. Having Domino do spam filtering should be your last resort, not your primary strategy.
The thing I like about client-side filtering, though, is that the user is responsible for deciding what is spam. Of course it is also the user that is responsible for getting spam....everytime you put your email address on the web (either posted on one of your pages or on an order form), you are a target. Also anytime your real domain name is on the web, you get whammied as well. I get many spams just because "www.waresource.com" is on the web...they generate fake names until there is a hit. And then I get a zillion because I am listed as the domain admin in whois. So there is another batch. My point is that even though I am employing client-side filtering, I still get spammed to death. So there is nothing you or anyone can do to stop it completely short of unplugging from the net. OR, use private tools that work by invitation only (such as Groove) for communication with outside companies.
The latest proposed legislation in the US won't help either...
In general, I wouldn't recommend the Anti-Spam agent any more. Once you get more than a few people using it on a server and more than a hundred spammers listed, server performance really takes a hit. Lotus should build something into the base Notes product for client-side action. That is just because of the time it takes for the before mail agent runs and soon the server will be overwhelmed with lots of users filtering lots of names. So while it works for a few people, you will cripple your server (specifically mail routing) if everyone is using it. I have changed the code in my personal copy to run on a schedule every 10 minutes on just new documents instead of as a "before new mail arrives" agent. This reduces the load considerably.
But I am surprised by Notes R6 not having expanded rules to handle SPAM more easily (it was mentioned in an article I read that it was a future consideration). They are talking about client-side SPAM filtering in the post 6.0 timeframe...but that was just one of those interviews with a product manager and doesn't mean it will surely happen.
See: http://www.crn.com/Sections/BreakingNews/dailyarchives.asp?ArticleID=34389
All I know is that we could continue to cobble together LotusScript solutions that only work moderately well, while the real solution must happen much deeper in the product code (and only Lotus can do this).
If you still want to use the agent, here are some possible trouble-shooting tips:
- I assume that you have the ACL right to Create agents (of course you do....).
- you need to copy all the views, forms, and agents to your mail database (you need Designer to do this).
- One thing is that the agent must be running on your mail server as specified in your Person document (or is that for after mail arrives agents? see kbase 158768)
- You will need to increase the time before mail agents run (kbase 174701) as you increase the number of spammers.
- Of course, if DESIGNER runs and the agent is not marked as "do not refresh", then it will be disabled (141036).
Sorry I don't have a happy answer for you! This is a very technically challenging problem, spam.
Best of luck in your anti-spam efforts!
Kent Kurchak
http://www.wareSource.com
(508) 839-6916 |
Go back |
| |
|