[Spambayes] Moving a SpamBayes Database between platforms?

[Mark Hammond]

Hi Hadar,

> I am running the Outlook plugin, and so far getting "mixed"
> results. I think that this is due to the fact that I collect
> email for _many_ addresses through one POP3 account.
> Meaning, the other accounts are all aliases or forwards into
> the one pop account.
>
> As an example, if I get three identical spam emails sent to
> me at three distinct addresses, one is caught by SpamBayes,
> and the other two are not. They're rarely tagged as
> "Possible Spam" either.
>
> Anyway, that's with training on roughly 3000 spam and 2700
> ham messages.

It would be interesting to see the "Spam Clues" for these messages.  These
clues should tell us why we are getting that behaviour.  It may simply be
that to date, these other addresses have not collected enough spam to tip
the clues over that edge - for example, if one address is "hadar at foo.com",
and there are only examples of ham with that address specified, that address
may be enough to convince the system it is ham.

An interesting experiment would be to have an option to ignore the "To"
address.  For example, every single mail to "skippinet.com.au" goes to my
mailbox (ie, mhammond@, foo@, death-to-spammers@, etc).  Thus, I have alot
of spam in my mailbox to apparently "random" addresses at this domain.  For
me, SpamBayes does a good enough job in this area - but experiments here may
benefit everyone.

> Tim Peters kindly pointed me to the FAQ and this list today,
> and I saw the answer on how to hook up the filter to
> procmail, which is exactly what I would like to do. However,
> my "twist" is this:
>
> I would like to run all of the training on my client, via
> the Outlook plugin. Then, on occasion I would like to
> "upload the database" to my server, and use the "static"
> database to filter via procmail.
>
> The main reason is so that I can filter what gets sent to my
> Blackberry, while continuing to more iteratively train on my
> laptop via Outlook.
>
> Is this straightforward to do, meaning, are the databases
> interoperable? Outlook 2000 running on Windows 2000
> Professional, with Procmail running on RH Linux 8.0 on the
> server side.

It *should* be possible, but you are entering uncharted waters.  Apart from
the database issue, a problem may be that the format of the mail once seen
by Outlook is different than seen by procmail - thus, the exact same mail
may create a different set of tokens depending on where it is seen.

ie, I guess I am saying "I don't know, but would love to hear the answer
once someone actually tries it" :)

Regards,

Mark.