Since a few days, I get my private mails delivered into my Notes mailbox, so that I can process my private mails as well as my normal "
company" mails with eProductivity. But most of this article is not specific to eProductivity, you can do this with the normal mailfile as well.
There are great instructions how to setup notes to get the mails from a POP3 server on the "Lotus Notes 8.x Tips" blog. If you get mails from both a domino server and a POP3 server, I recommend to first create a new location and switch to it and then run the described wizard... I had to recreate my normal (-> mail goes out via Domino Server) online setup afterwards.
Anyway: this works really well apart from one point: up to now I read mail my private mail in Thunderbird and used ThunderBayes to filter out spam. Notes, unfortunately does not have a spam filter, but this is how you can work around this limitation. Another limitation is, that notes does not process mailrules for POP3 mails, but there is an OpenNTF project for that. So here is the setup to fight spam in POP3 mails delivered to the Notes client:
1. Setup SpamBayes
2. Configure Notes to use SpamBayes
3. Setup POP3 mailrules
4. Configure a Mailrules to filter Spam
(5. Train the filter)
Setup Spambayes
I use SpamBayes, which is a baysian/learning spam filter which acts as a proxy beween the POP3 server and your mailclient and classifies the mail as ham (=good), unsure or spam (=bad :-) in some custom mail-header. Just download the windows version and install
(make sure it is started automatically as a service!). Do a restart of windows, to make sure you have the service running... Afterwards point you webbrowser to http://localhost:8880/config and configure SpamBayes to connect to your current POP3 server and a local port number (the rest of the options are ok for now :-). You don't need to input your username/password into SpamBayes.
Configure Notes to use SpamBayes
Now configure your Notes POP3 mailsetup as you would before, just with "localhost" and the local port number and Notes would get the mail through SpamBayes. If you have already configured it: the config is in the personal adressbook under "Advanced -> Accounts". My final setup looks like this:
Now you should already get mails delivered into your mailaccount which have the new mailheader: Open a new mail, goto "View -> Show -> Page Source" and you find something like this:
Setup POP3 mailrules
No we need something to filter on this. Go to the POP Mail Rules OpenNTF project and get the latest release. It has a nice Install instruction, but for the impatient: copy the rules.dll next to your notes.ini file and add a "EXTMGR_ADDINS=rules.dll" line to your notes.ini (if there is already one, read te instructions...). Restart Notes and your POP3 delivered Mails will be filtered by your mailrules.
Configure a mailrule to filter Spam
Now we have only one problem: how to get a Mailrule to filter on the "X-Spambayes-Classification:" Header? This is also done by a trick: the new extention has some extra capabilities, which can be accessed when you build a rule based on the BCC field: instead of you adding the string you would filter, you add "header:string" and the rule engine will base the filter on that header instead of the BCC field or better that field in the final mail document in your mailfile. There is unortunately a difference beween header names and fields: notes converts '-' in headernames into '_' in fieldnames. So if we want to filter by Mail with "X-Spambayes-Classification: spam", we need to add a BCC rule with a "contains"-String of "X_Spambayes_Classification:spam" (and move that mail into the junk folder (in eProductivity, Mailrules are under E-Mail -> Tools -> Rules):
And that was it: from now on, every classified spam mail should be filtered into the Junk folder.
Train the filter
On problem remains: how to train the filter. This is best done via the webinterface at http://localhost:8880/review: just classify the daily mails there and train the filter with that data. After a few days, it should get better and better until you get less and less false classified mails. For this to work, you need to cache messages in SpamBayes (see SpamBayes webinterface -> Configuration -> Storage options). The other option, to send messages to "spam@localhost" or "ham@localhost" does not work, as Notes does not send every old mailheader. Also, we haven't configured notes to send maisl through spambayes.
Happy emailing!
