Re: Pyzor usage by University?

[Vincent F. <vb...@uc...>]

On 01/22/2013 09:16 AM, Benny Pedersen wrote:
>
> We are currently just doing check per the example given.
> means only pyzor CLIENT ?
>
>
Yes.

We have a range of other measure in place.

We run DCC here, we get feeds through Rhyolite.
We also run various blacklists, and we host local
copies of that data on servers running rbldnsd.

So we have some experience and infrastructure
that could be leveraged, but as yet I know fairly
little about running Pyzor other than we are now
running it in client mode on our mail routers.

I would think a local-only pyzord would be inferior
to using data that is shared from other sites, who might
see a given spam before we do.  But I suppose it's no
different than running Bayes which we do local-only.

Any thoughts about best direction are welcome.