[Spambayes] Client/server model

[Guido van Rossum]

> > > I'd want the server to do tokenization for consistency reasons.
> > > Particularly if you are also spam filtering news articles and not
> > > just e-mail messages.
> > 
> > I don't understand this.
> 
> So that everybody tokenizes the incoming messages in the same way,
> particularly the same way as that used earlier during training.

The hammie-client approach has a separate client program that's
invoked each time, and that takes care of the uniform parsing.

> Also, I'd have the server keep track of spam from other sources,
> such as UseNet news.  Is there anywere else where spam messages
> show up that might need to be included, or is it just mail and
> news?

Not that I know of.

--Guido van Rossum (home page: http://www.python.org/~guido/)