[Spambayes] Deployment time

[Derek Simkowiak]

> 	I'll be reporting my experiences here to the list [...]

	First note: The filter-mode header of "X-Hammie-Disposition" seems
inappropriate, and let me explain why.

	I, the sysadmin, know what hammie.py is.  I know that I installed
it, and I know that it filters for spam.  I know that it is part of the
SpamBayes project, and that the header is inserted into spam-like
messages.

	However, someone else looking at the "X-Hammie-Disposition" header
out of context would not know at all what that header means, what to do
with it, or that they can filter on it for classifying spam.  A Google
search for "Hammie" does not give any results relating to the SpamBayes
project, and even worse, a search for "X-Hammie-Disposition"  gives no
results at all.

	It would be much more useful to use a header that can be
recognized for what it is, without having to be one of the rare
individuals who knows what "hammie.py" is.

	I suggest something like

X-SpamBayes-Disposition  [or]
X-Spam-Disposition  [or]
X-Spamfilter-Disposition

	...or, better yet, to stick with the conventions that SpamAssassin
has used.  This would be easiest on endusers and helpdesks, since setting
up filters for a SpamBayes installation would be the same as doing it for
a SpamAssassin installation.  Mobile users with email accounts in both
kinds of domain would only need one set of filter rules.  That, and the
SpamAssassin headers are pretty intuitive.

	SpamAssassin has slightly different semantics for its headers, but
it will be trivial to implement them in hammie.py.  If the maintainer(s)
are in favor of this approach, I can submit a patch in a couple of weeks.

	For reference, you can see how SpamAssassin tags spams at the
following URL:

http://spamassassin.taint.org/doc/spamassassin.html#tagging:




Thanks,
Derek Simkowiak