General questions regarding MS, SpamAssassin stuff

[Jason Williams]

>MTA: less resources, less control
>MS: medium resources, medium control
>SA: most resources, most control
>
>There's more in the FAQ.

Hmm. Each have their advantages and disadvantages. I'm sure there is some
sort of combo that can be used to take advantage of them.
I'll check out the FAQ.


>CA eTrust is cheap if you want another one.

I appreciate that. I was going to look at that one and F-Prot.


>Are you talking about manual training of the db? Depending on your
>e-mail system it can be tricky, Exchange (no surprise) is the worst. I
>don't train my db manually and have very good Bayes results anyway. It
>learns itself from very low and very high scoring messages. Start
>running Bayes and if you don't think it's accurate enough, then add
>manual training.

No exchange server here. :) I was reading over some of the SpamAssassin
docs and it mentioned keeping a copy of email it tagged as spam to a
certain directory. You can then run a command (ham I believe) that will
have SA read the emails in the directories and make itself 'learn' sort a
speak. I'll dig farther into that.

Bayes is something that I can turn on and let it go i'm guessing? I'll try
and find more about that as well.

I appreciate the help.

Jason